THE DEVELOPMENT OF CALCIUM CARBONATE - IRON OXIDE PROTECTIVE COATINGS ON IRON rjahmoud 0 . Ab du 11 ah. AN ABSTRACT Submitted to the College of Graduate Studies of Michigan State University of Agriculture and Applied Science In partial fulfillment of the requirements for the degree of Doctor of Philosophy Depart rent of Chemical Engineering 19 3 C THE DEVELOP] [TNT 0- CALC ILL I CARBONATE - IHON OXIDE PROTECTIVE COATINGS ON IRON A3STRACT The development of useful protective coatings for pre­ vention of corrosion in water systems has been under inves­ tigation using both static and dynamic tests. All work has been with water containing calcium hydroxide, carbon dioxide, oxygen, and nitrogen. Review of the literature indicated that a good under­ standing of the effects of pH, temperature, salinity, and other factors on the equilibrium of calcium carbonate in water has been reached. However the nature of the scale and the factors influencing its formation have not been com­ pletely established. It has been the ob'ect of this research project to de­ velop ’uniform coatings in a dense impermeable form that will provide high anti-corrosion protection, and to investigate and evaluate the effect of certain physical and chemical en­ vironmental conditions on the formation of these coatings. Petrographic methods were used to identify the minerals developed on the specimens. Coating materials developed on cast iron specimens were largely limonite. percent calcite was commonly present. Live to forty Siderite and magnetite were usually observed covered by limonite. Uith stainless steel specimens, only calcite was found* Polarization studies Indicated that CaCO^ acted pri­ marily as a cathodic Inhibitor# The action of calcium car­ bonate in developing good protective films seemed to lie in forming a physical mixture with corrosion products. Under favorable conditions the mixture bonded well to the cast ir on specimens and was hard and relatively tough. Better protection and better Bonded, harder, and tougher coatings resulted from solutions containing colloi­ dal CaCOg than from Identical solutions of the same pH and hardness with no colloids present. Colloidal CaCOg was found to have a positive charge In the pH range 6 - 1 1 may be represented as (CaCO3 ) Ca++JoiI"* and Calcium carbonate crystals In suspension did not have a favorable effect on formation of coatings from s rcsrsaturated solutions. Relatively high flow velocities of about 2 feet per second were desirable In the formation of hard durable coatings. Static tests generally produced soft coatings. A saturated level of oxygen was found to be optimum for de­ veloping good coatings under the conditions conducted In this study. A high momentary excess of CaCCg led to the formation of chalky soft coatings. Low momentary excesses and high saturation excesses of CaCQg led to the formation of hav*''‘!, Thus, it has been established that when a water is sa­ turated with dissolved oxygen, Is of pH 8.2 to 8.7, contains a momentary excess of 2.5 to 5*5 ppm CaCOg In the presence of colloidal CaCO^, and flows at a velocity of about two feet per second, a uniform, hard and well bonded coating can be developed on cast iron specimens in one dayTs time at room temperature. THE DEVELOPMENT OF CALCIUM CARBONATE - IRON OXIDE PROTECTI RE COATINGS ON IRON By Mahmoud 0. Abdullah A THESIS Submitted to the College of Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy Denartment of Chemical Engineering 19 08 Approved ProQuest Number: 10008565 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest. ProQuest 10008565 Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code Microform Edition © ProQuest LLC. ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 4 8 1 0 6 - 1346 ACIQiOULEDGhENT This thesis was only possible with the sincerely ap­ preciated support of a N.I.H. grant. The author wishes to express his sincere appreciation to Dr. R.F. McCauley and Dr. C.F. G-urnham for their valu­ able guidance and assistance in connection with this thesis. Thanks are also extended to Dr. H. Storehouse and Mr. C. Star for their help In the analyses of the coatings de­ veloped. TABLE OF CONTENTS Page INTRODUCTION.............................................. I LITERATURE REVIEW........................................ .3 TUEORETICAL C0N3JDEEATIONS................................9 A. B. C. D. E. Electrochemical Theory of Corrosion ofIron.......... 9 Typical Corrosion Cell on the Surface of Submerged Iron......................................13 The Action of Inhibitors................ ....... ...,l5 Polarization and the Rate of Corrosion..............19 Different Indexes Used for Calcium Carbonate 2k Equilibrium......... 1- Langelier Saturation Index...................... 2k 2- Ryznar Stability Index.......................... 2? 3- Saturation Excess......... 27 k“ Momentary Excess .......... .....28 EXPERIMENTAL APPARATUS AND MATERIAL...... A. B. C. D. Static Test Unit*..*................. ............... 29 Dynamic Test Unit with Rec irculatedWater............31 De-ionizing Units................. 38. Electrical Measurement '"nit..*.......... ....36 1- Potentiometer. ..... 36 2- Batteries ...... .36 3- Standard Cell 3q k- Galvanometer* .......... *....39 39 5- DC Power Supply.*.............. EX PER IMENTAL PROCEDURE A. B. C. 29 ....... * ..*.'■ 0 Static Tests........ . ...IQ Dynamic Tests with Recirculated Water*............. kl specific Investimations, . .. .. «, ... ,. .. ... .* . . .. ..,.in 1- Potential Measurements * ,'•3 2- Preparation of Colloidal Calcium Carbonate......' 3- Determination of ElecJ-rim a l Charge of Colloids ..' ? k- Analysis of Specimen Coatings After the Tests...' C .... 1 5- Routine Analytical Determinations Page CONDITIONS OF THE TESTS AND ANALYSIS OF THE COATINOS DEVELOPED.................................................50 DISCUSSION OF RESULTS.............. . .61 Mineral Contents of Developed Coatings •.......01 Polarization Studies.......... , ........ 62 Formation of Colloidal CaCOQ.................... .*.68 Electrical Charge of Colloidal CaCOo.............. .7J-JEffect of Colloidal CaCO^ on Coatings Developed....76 Effect of Momentary Excess on Coating Development••85 Effect of the Age of Colloidal Particles on the Rate of Deposition of CaCO^ and Coating Development......... 93 Effect of Dissolved Oxygen Levels on Coatings 98 Developed............... Effect of Specimen Surface Conditions on Coatings Developed. ..................... 107 Effect of Iron Oxide on Coating Development on Stainless Steel Specimens......................... 112 Static Tests VS. Dynamic Tests....... llh CONCLUSIONS............................................. 116 RECOMMENDATIONS.......... 118 BIBLIOGRAPHY............. 120 LIST OF FIGURES Page Figure 1 Schematic Diagram of Corrosion Cell on Surface of Submerged Iron....................... lR 2 Typical Polarization Curves..................... 20 3 Types of Corrosion Control Curves....... Ur Equipment for Dynamic Tests 3 Equipment for Static Tests...................... 30 6 Flow Diagram of Dynamic Tests with Recirculated 32 Water........... 7 Test Cell Construction.......................... 33 8 Large Two-Bed lon-Exchanger 9 Potentiometer Circuit Diagram....... ........... 37 23 ,. .28 ........ 35 10 Wiring Diagram of Electrical Equipment....... .Jgli 11 Polarisation Curves of Dynamic Test 1*.......... PR 12 Polarization Curves of Dynamic Test 13 Polarization Curves of Dynamic Test 3 ll- Curves Showing the Relationship of Momentary Excess and Saturation Excess Values to the pF Levels of Supersaturated Solutions at 'filch Colloidal CaC03 Formed. ..... 2 .......... 65 ........ 66 .72 15 Schematic Diagram of Colloidal CaCO^ PartIcle...77 16 Potential-Time Curves showing vhe Effect of Colloidal CaCOg on Coatings Developed In Dynamic r’ests h and 3 .. .......... q Figure Page 17 Potential-Time Curves Showing the Effect of Momentary Excess Levels on Coatings Developed in Dynamic Tests 7 , 8, 11, and 12............... 89 18 Potential-Time Curves Showing the Effect of Fate of Deposition of CaCO^ on Coatings Developed in Dynamic Tests 15 and 19........... .95 19 Potential-Time Curves Showing the Effect of Rate of Deposition of TaCOy on Coatings Developed in Dynamic Tests 12 and 13............97 20 Potential-Time Curves Showing the Effect of Dissolved Oxygen Levels on Coatings Developed in Dynamic Tests 20, 21, 22, and 21g............ 103 21 Potential-Time Curves Showing the Effect of Dissolved Oxygen Levels on Coatings Develo :sd in Dynamic Tests 17, 18, and 19................ 105 22 Potential-Time Curves Showing the Effect of Cast Iron Specimen Surface on Coatings Developed in Dynamic Test 11........ . . 23 109 Potential-Time Curves E::owing the Effect of Stainless Steel Specimen surface on Coatings Developed in Dynamic Tests 12 and 13..........,111 LIST OF TABLES Table Page 1 Conditions and Analysis of Coatings of Static Tests . .. . ....... . . .......... .. .......... 52 2 Conditions and Analysis of Coatings Developed of Dynamic Tests...................... 3 Percent Compositions of Coatings Developed in Dynamic Tests 1, Z f and .................. 67 Data of pH and Hardness of Supersaturated Solutions for Formation of Colloidal CaCO^...... 70 Effect of Colloidal CaC03 on 'Height Losses and Weight Gains for ^ome Static Test Specimens.80 Effect of Momentary Excess Levels on Coatings Developed In Some Dynamic Tests................. 87 Effect of Momentary Excess Levels on Coatings Developed on Cast Iron specimens for Some Static Tests ............. 90 Effect of Momentary Excess Levels on Coatings Developed on Stainless Steel Specimens for Some Static Tests....... ...92 Effect of Dissolved Oxygen Levels on Coatings Developed for Some DynamicTests.................99 INTRODUCTION All water systems -undergo corrosion, generally at slow but continuous rates. In some systems waters are so "aggres sive" as to seriously damage pipes and appurtenances within a few years. In other instances, tuberculation continues over long periods, gradually decreasing the carrying capa­ city of the system and resulting in various degrees of "red water." Losses from corrosion are of major importance in both municipal and industrial water works. In addition to direct economic losses, the water works industry suffers from over-design of structures in anticipation of corrosion, reduced capacities of pipelines, increased hazard of struc­ tural failure during peak or fire demands, and impairment of water quality by corrosion products entering solution. For these reasons the development of means for reducing corrosion in water systems Is of prime importance. Natural protective coatings, usually of calcium carbonate or mix­ tures of calcium carbonate and corrosion products, are now widely used in water works practice for prevention of exces­ sive corrosion and for overcoming "red water" problems. No satisfactory procedure so far exists, however, for developin satisfactory protective coatings under all normal operating conditions. Coatings are in some instances dense, hard, and 2 uniform; while in other instances they are soft, rough, and permeable• It has been the object of this thesis to learn how the coatings can be uniformly laid down in a dense impermeable form that provides high anti-corrosion protection without damage to hot and cold distribution systems. To accomplish the objectives of this thesis problem, the mechanism of lay­ ing down these desirable protective coatings has been studied and an investigation of the effect of certain physical and chemical environmental conditions on their formation has been made. Petrographic methods have been used to identify the materials which formed the scales -under various conditions. 3 LITERATURE REVIEW Anhydrous calcium carbonate occurs in nature in at least three known forms under various conditions of preci­ pitation (1)* Of these three forms, rhombohedral or cal- cite is the most stable at ordinary temperatures. Orthorhom- bic aragonite is only slightly stable and, In water, tends to transform Into calcite at an extremely low rate at ordi­ nary temperatures and pressures. The third form, vaterite, is unstable and changes into either aragonite or calcite. The earlj use of calcium carbonate scale for inhibiting corrosion was necessarily on an empirical basis. practical and useful tests mans and collaborators In time, (such as the marble test by Till­ (2 )) were developed to determine the degree of supersaturation or undersaturation of a water. These methods helped to predict the tendency toward deposi­ tion or removal of scale. The development of the "saturation index" by Langelier (3 ) based upon principles of physical chemistry which will be dealt with later under theoretical considera­ tions. The "saturation index" is the measure of the driving force tending to bring a water to stability. A positive in­ dex Indicates that scale will be laid down, a negative Index that scale will oe dissolved. A water may be adjusted tc a positive scale forming index by adding lime and/or sodium carbonate. Such adjustment is easily accomplished by m e ­ thods suggested by Moore ([|.) and others. Enslow (5) has devised a "continuous stability indicator" to measure the degree of supersaturation or undersaturation of a water. Refinements in computation of the "saturation index" have resulted from further studies by Langelier (6 ) and others. McKinney (7) has discussed the calculation of equi­ libria In dilute water solutions and has pointed out the ef­ fect of alkalinity on changes In pH that might be expected due to temperature increases. Larson and Buswell (8 ) have discussed the effect of salinity on Ionization constants, thus increasing accuracy in determining "saturation indexes" at different temperatures. Ryznar (9 ) Las proposed an empirical "stability index," which will also be discussed later under theoretical consi­ derations, with a view to providing an indication of the de­ gree of scaling that can be anticipated. It has also been suggested by Dye (10) and others that the deficit or excess of carbon dioxide in a given alkalinity-calcium-pH system might be used as an index of scaling. From the preceding information It can be seen that a good understanding of the effect of pH, temperature, salinity, and other factors upon the equilibrium of calcium carbonate in water has been reached. However, the nature of the scale 5 and the manner in which it is formed are far from being completely understood at the present time. It lias been shown by Larson and King (11), Raistrick (12), and Evan (13) that calcium carbonate can be deposited from calcium bicarbonate solution in an electrochemical process at the cathode, The precipitated calcium carbonate is directly proportional to the current density. According to Evan (lip) and Haase (15) , a heterogeneous layer of rust and calcium carbonate gives a better protec­ tion from corrosive attack than does a layer of rust alone or a layer of calcium carbonate alone. Evan (lip) has expressed the opinion that rusting iron yields ferrous oxide or hydroxide which is then oxidized to the less soluble ferric oxide. Additional ferrous oxide then diffuses outward through the ferric oxide and iron continues to move into solution in spite of the coating of ferric oxide. Calcium carbonate in the presence of oxygen interacts with iron salts to form a clinging ferric oxide rust. If oxygen is present in large amounts, the rust is formed very close to the metal and ferrous salts and cal­ cium carbonate Interact to yield ferrous carbonate which is then oxidized. In the absence of oxygen, magnetite ap­ pears; this is loose and not protective. The data of Strum (16) on cast iron specimens exposed to oxygen-saturated water of varying hardness and pH 6 indicates that a high proportion of calcium carbonate in the film is more desirable in retarding corrosion than an increase in film thickness* Coatings developed were found to contain a much higher percentage of calcium carbonate in the layers closest to the iron surface than the exterior layers which were predominately iron oxide* Baylis (17) has shown that ferrous carbonate is pro­ duced in varying quantities as a product of corrosion if the carbonic acid is present either as free or half-bound carbon dioxide (CO2 and HCO3 ) • Reaction takes place either directly with metallic iron or with iron oxides or hydrox­ ides. Baylis' tests also showed that ferrous carbonate in the absence of oxygen is practically insoluble at a pH and alkalinity close to the calcium carbonate saturation curve, but that the solubility increases very rapidly if the pH is decreased below the level necessary for calcium carbonate equilibrium. Baylis has stated that ferrous carbonate aids in protecting iron surfaces where the pH and alkalinity are above calcium carbonate equilibrium, especially if the fer­ rous carbonate is overlaid by some kind of a coating that protects it from dissolved oxygen. The relationship between the formation of coatings and the dissolved oxygen level in water has not been completely established. American water works practice is generally to reduce dissolved oxygen to the hr nrm possible level since 7 oxygen is the primary factor controlling the corrosion rate. On the other hand, European workers, as reviewed by Evan (lip), intentionally aerate the water before it enters the pipe line so that the dissolved oxygen concentration will be increased. Their argument is that corrosion resulting from the oxygen content of the water is desirable for pro­ moting and accelerating deposition of a calcium carbonate protective coating on the metal. According to Schikorr (1 8 ), Haupt (19 )9 and others, an oxygen concentration of at least 6 mg per liter is necessary to produce a protective coating• Gamp (20) in his thermodynamic treatment of an elec­ trochemical corrosion has shown that ferrous hydroxide, Fe(0H)2* will not form on iron unless the pH value of the solution exceeds 9. Evan (lij.) has reviewed the work of a number of authors who have reported on different species of rust developed in electrochemical corrosion. Quoting Schikorr (21), Evan has stated that the oxidation of ferrous hydroxide to hydrated ferric oxide FeO(OH) results in goethite if the oxidation is rapid and produces the mineral lepidocrocite if slow oxidation takes place. In the formation of lepidocrocite, a ferrous-ferrite material is first developed which in turn yields the hydrated ferric oxide. G-irard (22) has suggested that ferrous hydroxide is 8 first formed on the metal face of corroding iron and that exposure to oxygen changes the deposit to the ferric con­ dition, The ferric material in turn reacts with ferrous hydroxide to form green hydrated magnetite, Pe30[j_«H20 which acts as an intermediate body. The diffusion of oxygen to the green body yields either goethite or lepidocrocite. If oxygen Is deficient, the green Intermediate undergoes dehy­ dration to black inert magnetite. Thus, stratified layers of different iron oxides are often formed in the electro­ chemical corrosion of iron. 9 THEORETICAL CONSIDERATIONS A. Electrochemical Theory of Corrosion of Iron The electrochemical theory of corrosion proposed in 1903 by W. R. Whitney (23) has been widely accepted by stu­ dents of this phenomena. The theory has been discussed by many prominent authors in the field of corrosion, Including Uhlig (2L!_) , Speller (25), Evan (lly), Eliassen and co-wor­ kers (26), and others. The fundamental mechanisms involved in this theory can be described in the following manner: 1. The process of electrochemical corrosion requires the presence of anodic and cathodic areas. anodic area (electronegative) The is the area over which the metal is attacked (oxidized), and the cathodic area (electropositive) is that at which some substance from the environment is reduced. If the process is to continue, an electric current must flow between these areas through the environ­ ment (usually an aqueous solution) and between these areas in the metal. This electric current is carried oy the electrons through the metal and Is carried jointly by cations and anions, which mi­ grate In opposite directions through the solution. 10 The combination of anode area, cathode area, and aqueous solution constitutes a small galvanic cell* Cells of this type may be set up on a single metal­ lic surface or between dissimilar metals* 2*. A flow of electricity results from the loss of two electrons from each iron atom leaving the crystal lattice of the metal as ferrous ion and entering the solution at the anode* Fe° = Pe++ + 2e (1) 3* Water partly dissociates into hydrogen ions and hy­ droxyl ions (one molecule in every 5 5 5 *000,000 dis­ sociates into its constituent H+ and 0H“ ions) H20 = H+ + OH- (2) Ferrous ions react with hydroxyl ions to form fer­ rous hydroxide, Fe( 0H)2 , which is soluble in water to the extent of 7*0 ppm at 20° C. Fe++ + 2(OH)“ = Fe(0H)2 (3) In the presence of oxygen, ferrous hydroxide is oxidized to ferric hydroxide, FeCOH)^, or hydrous ferric oxide, commonly recognized as rust. IpFe (OH) 2 + 2H20 + O 2 - lpFe(0H )3 Ig. At the cathodic area the electrons (I4.) areaccepted by the hydrogen ions to form elementary hydrogen atoms . 2H+ + 2e = 2H° (5) 11 This removal of hydrogen ions causes an increase in the concentration of hydroxyl ions in the area, and results in the production of alkaline conditions in the vicinity of the cathode, 5* If the hydrogen thus formed remains on the surface the area becomes polarized by the counter potential produced. This retards the flow of electrons and consequently the solution of iron. 6 . In the majority of cases, however, the areas are depolarized, allowing the action to continue. Had­ ley (2 7 ) has enumerated five depolarization proces­ ses which may govern the rate of hydrogen removal from the cathode. a. Reaction of dissolved oxygen with nascent hydro­ gen to form water. 2H° + i02 =:H20 (6 ) b. Agitation causing the removal or sweeping off of hydrogen as gas bubbles. 2H° —-H2 (7) c. Direct chemical combination with electrolyte or salt. For instance, ferric ions may be reduced to ferrous ions by the hydrogen or by the elec­ trons . 2FeH"f+ + 2H° = 2Fe++ + 2H+ (8 ) 2Fe+++ + 2e = 2Fe++ (9) 12 d. Reaction in the metabolic processes of certain anaerobic bacteria. e. Combination with the products of microbiological metabolic processes* The first two of the above processes are well known and may be observed in varying degrees in water corrosion processes. In acid solutions, usually of pH below 5, cathodic film is removed mainly as bub­ bles of gas. The hydrogen ion pressure is suffi­ cient to overcome the hydrogen overvoltage at the metal surface, causing marked evolution of hydrogen gas. This accounts directly for the fact that cor­ rosion is generally more rapid in acid solutions and less rapid in alkaline solutions. In neutral or alkaline solutions, the amount of gaseous hydro­ gen is very small compared with the amount of h y ­ drogen destroyed by oxidation. Thus, in the pH range commonly encountered in natural waters, de­ struction of the hydrogen film is mainly governed by depolarization with dissolved oxygen. Since the cathodic reaction controls the rate of corrosion process, a3 will be explained later in this section, the depolarization reaction and, therefore, the rate of oxygen supply to the cathodic regions of the iron surface is the factor governing the rate of corrosion in natural waters. B. Typical Corrosion Cell on the Surface of Submerged Iron Figure I shows schematically a typical corrosion cell on the surface of a submerged piece of iron. The anode and cathode of this cell are short circuited by the body of the metal. The surface of a large section of metal might be covered by many such corrosion cells. Formation of anodes and cathodes on a submerged metal are due mainly to (2 0 , 2 6 ) a. Difference in metal composition, b. Difference in electrolyte concentrations (forming concentration cells), c. Difference in aeration (the portion of the metal that has freer access to oxygen acts as the cathode, while the portion shielded from oxygen becomes the anode of the differential aeration cell), and d. Difference in temperatures or stresses. In the corrosion cell shown in Figure I, iron passes into solution at the anode area leaving electrons on the metal. These electrons pass through the metal to the cath­ ode area, where they may be removed from the iron through reaction with hydrogen ions in solution producing atomic hydrogen (Equation 5)$ or by one of many other half-cell reactions which depend on the substances present in the Ik 2e Anode Area IKON Figure 1 - Corrosion Cell on Surface of Submerged Iron 15 vicinity of the cathode and the relative energies governing half-cell reactions (26). Under certain conditions the hy­ drogen atoms tend to plate out on the metal and unless re­ moved will interfere with transfer of other electrons to solution. Within the pH range normally encountered in w a ­ ter works practice, it is generally considered, that remo­ val of hydrogen film from the cathode is by oxygen depolari­ zation (Equation 6 ). Wilson (28) has pointed out that anode reaction rate (Equation l) is much faster than the depolarization reac­ tion rate at the cathode (Equation 6 or 7)* It is obvious that if a process comprises two or more separate reactions, the rate of the process as a whole Is determined by the rate of the slowest of these reactions under the particular conditions. This is illustrated by the fact that though in the case of Iron an increase In the metal-ion concentration should lessen the rate of anode reaction (Equation 1), It is without appreciable effect on the over-all rate of corrosion. Thus, cathodic reactions, govern the rate of corrosion pro­ cess . C * The Action of Inhibitors Speller (25) has defined an Inhibitor as a chemical substance or mixture which, when added to an environment, 16 usually In small concentration, effectively decreases cor­ rosion, The protective action of Inhibitors is due mainly to the formation of films on the metal surface. films in turn retard the corrosion reactions. These Film forma­ tion may occur preferentially at either anodes or cathodes of corrosion cells, or it may be adsorbed generally over the entire surface of the metal. Evan (lip) has classified inhibitors into three classes: (1) Anodic inhibitors (2) Cathodic inhibitors (3) Adsorption inhibitors It should be noted that these classes are not clearly de­ fined since some inhibitors may act In more than one way (26) . Anode films may be formed through electrodeposition of negatively charged ions or colloidal particles on this electrode, through adsorption of chemicals on the metal surface, or through the production of an insoluble preci­ pitate by reaction between the iron entering solution and chemicals in solution. Anodic inhibitors reduce the rate of corrosion by restraining anodic reaction. The addition of a sufficient quantity of anodic inhibitor to a wate” could result In the formation of a continuous film over the anode areas which would be effectively separated from the water, thus stifling the anode reaction. 17 Although anodic inhibitors are very effective in reducing the over-all rate of corrosion, they are somewhat u n ­ desirable and even dangerous. As Evan (II4-) has pointed out, these inhibitors may intensify corrosion if used unwisely. Addition of insufficient anodic inhibitor to the corrosive medium would result in the formation of only partially pro­ tective film over the anode areas. It has been pointed out previously that the rate of corrosion of iron In water is generally governed by the rate at which cathodic reactions take place. Therefore, partial covering of the anodes has little effect on the over-all rate at which metal passes into solution, and may result in Intensification of attack on the smaller remaining anodic areas with rapid failure of the structure by pitting. The use of anodic Inhibitors In municipal water systems, therefore, would require the clo­ sest possible expert control. Cathodic films may result from the electrodeposition of positively charged ions or colloidal particles 0:1 the cathode, from adsorption of chemicals on the metal surface, or from the insoluble precipitates formed on the metal by a reaction oetween chemicals in solution and the alkalinity produced at this electrode. Cathodic inhibitors reduce the rate of corrosion by Interfering with cathodic reactions. These Inhibitors, when used in proper concentrations (2-g), are less effective In reducing the over-all corros'on rate 18 than anodic inhibitors, due in part to the fact that a pro­ tective film is formed on the cathode. The film sets up a differential aeration cell, with the shielded cathode ten­ ding to become anodic to the exposed anode. in continual changes This results in the locations of anodes and cath­ odes while corrosion continues at a reduced rate (26). The protective film formed by a cathode inhibitor in­ terferes with the access of dissolved oxygen to the metal surface, and reduces the effective area of the cathode. Any reduction in the effective cathodic area must result in a corresponding reduction in the over-all corrosion rate. Cathodic inhibitors always render the corrosion less in­ tense, even if they do not completely arrest the loss of metal. If a cathodic inhibitor is added in an amount suf­ ficient to stop corrosion, the decrease in cathodic reac­ tion permits an extension of the anodic areas, and helps to make corrosion less intense (lip) • Thus, the intensity of corrosion Is diminished, under cathodic control, by redu­ cing the total corrosion and by increasing the area over which the attack is spread. Contrary to the use of an ano­ dic Inhibitor, miscalculation in the amount of chemicals required will not lead to intense attack at some point with a resulting failure of the structure. For this reason Evan (lip) has classified cathodic inhibitors as safe. Some organic substances appear to be adsorbed over the 19 entire metallic surface (llj.) . These compounds may be classed as adsorption inhibitors, and may interfere with either or both electrode reactions. D. Polarization and the Rate of Corrosion The action of corrosion inhibitors may be illustrated through polarization studies of either or both electrodes (IJ4., 2I+, 25, 26). It has been stated previously that, for the process of corrosion to continue, an electric current must flow between different areas of the metal surface. In general, when current flows between two electrodes, a spe­ cial opposition to current flow develops at the surface of the electrodes. This special opposition to current flow is termed polarization. Polarization of the electrode from which positive electricity leaves to enter the electrolyte (the anode) is termed anodic polarization, and that of the electrode to which positive electricity from the solution flows (the cathode) Is termed cathodic polarization. The effect of both types of polarization is to reduce the ef­ fective difference in potential between the local anodes and local cathodes. Thus, polarization reduces current flow and consequently retards corrosion. Figure 2 shows a typical pair of polarization curves, indicating the potentials of the anode and cathode of a Electrodes Active Potentials Noble 20 Corros ion Potential Corrosion Current t Electrical Current Flow Figure 2 - Typical Polarization Curves 21 corrosion cell at various values of current flow between the electrodes• The points at which the curves intercept the ordinate axis represent the open circuit potentials of the electrodes (no current flowing in the corrosion cell)* As the current flow through a corrosion cell increases, the potentials of the anode and cathode tend to approach a com­ mon value* Polarization of both electrodes would result in a shift in the anode potential in the cathodic direction and a shift in the cathode potential in the anodic direc­ tion. Therefore, the potential difference between the electrodes in a corrosion cell through which current is flowing is usually less than the difference in the open circuit potentials (or potentials of electrodes with no current flowing). The intersection of the curves at point 1 gives the corrosion potential and the corrosion current with the electrode short circuited, which is usually the situation in practice with both electrodes located on sin­ gle piece of metal. It is to be noted that corrosion po­ tential represented by point 1 of Figure 2 has been deter­ mined for most of the dynamic tests of this study in order to follow the action of the inhibitor with time under vari­ ous conditions which will be discussed later. The rate of corrosion can be controlled by the rate of the cathode reaction, the anode reaction, or both. Alto­ gether, four types of control are possible in corrosion 22 cells. These are illustrated by Figure 3* The addition of anodic inhibitor to a corrosion system results in increased polarization of the anode, as illus­ trated in Figure 3a* The increase in the degree of polari­ zation is shown by an increase in slope of the anode line, indicating a greater change in anode potential for a given change in electrical current flow of the cell. In this case the corrosion is said to be under anodic control. The addition of the inhibitor results in a decrease in corro­ sion rate and a more noble corrosion potential. The addition of cathodic Inhibitor to a corrosion sys­ tem causes an increased polarization of the cathode. Fi­ gure 3b represents the condition under cathodic control, resulting in corresponding reduction in current flow and corrosion rate, and a less noble corrosion potential. Figure 3c shows polarization curves when the rate of corrosion is controlled by both electrodes (mixed control). Figure 3d represents the limiting case of resistance con­ trol, In which current flow between the electrodes has no effect on the potential of either of them. This situation would never occur in water works practice* The addition of an inhibitor to a corrosion system may have an added effect on polarization curves by causing a change in the open-circuit potential of either or both electrodes. Potential Noble ---- 23 Cathode Anode Current Flow b - Cathodic Control ►- a - Anodic Control Current Flow Potential Noble Cathode Anode Current Flow c - ilixed Control Current Flow & - Resistance Control Figure 3 * Types of Corrosion Control 2Ll In most municipal and industrial water systems, corro­ sion is primarily under cathodic control, as illustrated by Figure 3b* The rate of corrosion is governed by the degree of polarization of the cathode. E. Different Indexes Used for Calcium Carbonate Equilibrium 1. Langelier Saturation Index The basic reaction Involved In the reversible pipe scaling process can be written: CaC03 (solid) + H+ = Ca++ + HCO3 (10) At equilibrium, therefore, the product of the molal concen­ trations of calcium and bicarbonate divided by the molal concentrations of hydrogen Ion will remain constant, or i C a ^ M H C O H = K (11) Expressing each term as negative logarithms we can write for the pH at equilibrium pHeq. = pCa++ +■ PHCO3 - pK (13) By formulating PIICO3 in term of pH and total alkalini­ ty at all pH levels and by showing that the -pK term of equation (12) Is actually the difference between pK£ (the Ionization constant of the acid HCO3 ) and pKg (the solubi­ lity product constant for calcium carbonate), Langelier (3 > 6 ) has developed the following expression for pHeq #: pHs = (pK2 - pKs ) + pCa++ + pAlk. + log 1 + —y lj-S (13) Where pK^ and pKs respectively, are the negative logarithms of the second Ionization constant of carbonic acid and the solubility product constant of calcium carbonate, corrected for ion activity; pCa++ is the negative logarithm of the con­ centration of calcium ion in moles per liter; pAlk. is the negative logarithm of the titrable alkalinity in equivalents per liter; k 2 is the second ionization constant of carbonic acid; Hs is the hydrogen Ion concentration at a hypotheti­ cal saturation with calcium carbonate; and pHs Is the pH value corresponding to the above hydrogen ion concentra­ tion, that is the pH at which a water of given calcium con­ tent and alkalinity Is in equilibrium. Langelier also introduced the "saturation index” which is the algebraic difference between the actual pH of a sam­ ple of water and calculated pHs Saturation index = pH actual - pH saturation Saturation Index = log ■ 1 1 (j ^^ j - log QfTy = log Th^T (lip) Actually'this index, as seen from equation (15)» Is the lo­ garithm of the ratio of the hydrogen ion concentration which the sample must have if saturated composition) to its actual hydrogen (without change in Ion concentration. the Index is zero the sample is In equilibrium. A plus value for the saturation index indicates If 26 oversaturation and a tendency to deposit CaCO^; a minus v a ­ lue indicates undersaturation and a tendency to dissolve Ca003 • Saturation index varies with hot and cold water sys­ tems due to the difference in the rate of change of pH and the Langelier constant with temperature. A method for pre­ dicting this variation has been suggested by Powell, Bacon, and Lill (2 9 )• Dye (3 0 ) has also described a practical means for predicting pH values of a water at different tem­ peratures . Saturation index, according to Langelier, is an indi­ cation of directional tendency and of driving force, but it is not a measure of capacity, she amount of calcium car­ bonate deposited cannot be predicted from equilibrium data alone. Such deposition is dependent upon the relative rate of precipitation from solutions which are supersaturated with CaC(>3 . CaC03 has a strong tendency to remain in su­ persaturated solution and in the absence of crystallization nuclei oversaturated calcium carbonate solutions can be preserved for years. In general, more scale will be laid down at a low pH of saturation than a high pH of saturation. "Saturation index1' is often of benefit to water chemists in determining whether waters are scale-forming: or corrosive. Even in this respect :tsaturation index” is not always reliable, 27 because some waters with a positive index actually may be quite corrosive. This point has been observed and reported by Hoover (31) and others. 2 . Ryznar Stability Index Ryznar (9) has introduced an empirical expression, 2pHsaturation - pHactual, which he has designated "stabili­ ty index", to differentiate it from Langelierfs "saturation index." It has been claimed by Ryznar that "stability in­ dex" is not only an Index of CaCC>3 saturation, but is also of quantitative significance. Unlike the "saturation in­ dex," the "stability Index" Is positive for all waters. Experimenting with the formation of calcium carbonate scale on glass coil at 120-200° P., Ryznar has found Incrustation to take place within a two-hour test period at a stability Index of 3 and below. Ho deposits have been obtained at an index value of above 7.5. Ryznar*s "stability Index" has been widely used In waterworks practice. 3. Saturat ion Excess "Saturation excess" is the amount of calcium carbonate precipitation (in ppm) which must take place to bring a calcium-carbonate-bicarbonate-carbon dioxide system to equilibrium. It can be determined experimentally by the "marble" test, in which the water Is placed in contact with 28 powdered CaCO^, The decrease or increase in alkalinity is then a measure of "saturation excess11 or ’’saturation defi­ ciency11* ’Saturation excess” can also be determined graphi­ cally using methods developed by Langelier (6 ) and Dye (32). It, Momentary Excess Dye (32) has defined "momentary excess” as that frac­ tion of calcium and carbonate ions present (in ppm CaCOy) of an aqueous solution which is in excess of the solubility product constant of calcium carbonate. Dye also developed a graphical method for determining "momentary excess” based on the following equation: (Ca++ - X) (C0“ - X) = Kg 1010 (16) Where Ca++ and C05 are the initial concentrations of calcium and carbonate ions respectively, in terms of ppm of calcium carbonate; Kg is the solubility product constant for calcium carbonate corrected for temperature and dis­ solved solids; and X is the momentary excess in ppm of cal­ cium carbonate, ’’Momentary excess” has been extensively used in the work reported in this thesis because it is a quantitative measure of the driving force tending toward calcium carbonate deposition. 29 EXPERIMENTAL APPARATUS AND MATERIAL Work done on this project can be divided into two ca­ tegories: (A) static tests and (B) dynamic tests* A* Static Test Unit Each static test was conducted in a glass cylinder eight inches In diameter and eighteen inches high as shown in Figure 5* To each cylinder fourteen liters of de-ionized water was added, filling the cylinder to within one inch of the top. Two specimens of cast Iron or stainless steel were totally immersed in each cylinder by attaching each specimen to a glass rod by means of a nichrome wire and rubber band. The immersed portion of the nichrome wire was Insulated by a waterproof rubber tape. The two glass rods with the specimens attached were fixed vertically In each cylinder so that the specimens were parallel, facing each other with a spacing of four inches. The water in the cy­ linders was stirred throughout each test by bubbling air continuously. The air supply was first passed through a cotton bed to remove particulate matter and then through a scda-lime bed to remove G O g . 30 V.V ' • Figure Eq ui pment for D y n a m i c Tests Figure 5* Eq ui pment for Static Tests 31 D . Dynamic Test Unit with Recirculated Water Apparatus for dynamic tests with recirculated water is shown in Figures ip and 6 . a. fifty-five gallon polyethylene barrel was used as reservoir for the water, which was cir­ culated through the test cell by an all-rubber Universal Electric Company centrifugal pump* All parts of the pump in contact with the liquid were Hycar (synthetic rubber)* Construction of the test cell is shown in Figure 7* The cell was built of Lueite plastic to act as a housing for two metal specimens and to allow measurement of their potentials. A 3/^-inch hole was bored through the central axis and two parallel slots were machined, each at a dis­ tance of l/lp inch from the center. The two specimens were mounted parallel in the cell through the slots, separated by a distance of l/2 inch, in order to permit water flow parallel to their surfaces. holes were drilled and tapped on either side of the cell to connect each specimen with a platinum wire for potential measurements. Tapered Lucite inserts were used to reduce turbulence in the vicinity of the specimens, the width and thickness of the inserts being the same as that of the specimens. In some tests, where only one specimen was used, the test cell was of similar construction with only one slot located at the center of the 3/ip-Inch hole. One side only 00 O rO 1 00 r—f 00 +-» r— ) H in O 73 — /: 3ljr/A'£/A9 -«*-J N-^ I VO Nj 1*4 \J k 1° the cell was drilled and tapped for potential measurements. The cast iron or stainless steel specimens comprised the only metal in the system. All piping was Tygon tubing of 5/8-inch inside diameter and Van-Cor (unplasticized polyvinyl chloride) plastic fittings. The rate of flow was regulated by a clamp on the discharge side of the pump. As shown in the flow diagram of Figure 6 , a saturated calomel electrode was inserted in the line through a tee fitting, to act as a reference electrode during potential measure­ ments of the specimens. Saturated lime solution was added through the inlet tee fitting and carbon dioxide through the outlet fitting of the drum to adjust the water of the reser­ voir to the desired level of pH and hardness. Samples of the water were taken through a petcock fitting. C. De-ionizing Units All waters used for either static or dynamic tests were first passed through two de-ionizing units. The first unit was a large two-bed Lu-ilb Illco-way deionizer shown in Figure 8 . This unit was a strong-base ion-exchanger which could remove all ionizable solids including silica and COp. The water was then passed through a small mixed bed of Illco-Way research model de-ionizer. This small unit was also a strong-base ion-exchanger and could produce water 35 36 comparable to that produced by triple distillation* D. Electrical Measurement Unit 1* Potentiometer A Type K Leeds and Northrup Go* potentiometer was used during dynamic tests to measure the potential of the speci­ mens with reference to a standard saturated calomel elec­ trode* Figure 9 shows the potentiometer circuit and the electrical connections. The instrument consists of 15 five-ohm coils AD (adjusted to a high degree of accuracy) connected in series with an extended wire DB, the resistance of which is also five-ohms* A point M is arranged so that it can make contact between any two of the five-ohm coils and point M 1 is also arranged to make contact at any posi­ tion on the extended wire D B . Current from batteries flows through these resistances and by means of regulating rheo­ stat R is adjusted to one-fiftieth of an ampere. The po­ tential drop across any one of the coils AD is consequent­ ly one-tenth of a volt and across the extended wire DB. is 0.11 volt. D /■ placing the contact point IT1 at zero and mo­ ving the contact >1, the fall e;" Potential between II and II* may be varied by steps of one-tenth volt, from 0 to 1.3 vclts. The wire DB is divide.., to 1100 equal parts. moving the contact point M* alTr; the wire, the fall of By 37 ■VvW nwvviA O CHECK. 0<7 ^ R A / /£ ,£ V/0L75 S W trc # 6 t+A!N 51/£>£M*’ £'j STD CR14 Slt£)£Wfi£ T STdy criLp g x p t , O y. O _ ST£X CELL a E.M.F. Figure 9 # Potentiometer Circuit 36 potential between M and M f may be varied in infinitesimal steps , To use this variable fall of potential in making mea­ surements, convenient electromotive force is introduced in series with the galvanometer between the points M and M 1 and in opposition to the fall of potential along AB. The con­ tact points M and M* are then adjusted until the galvanome­ ter shows that no current Is flowing, and the value of the electromotive force can then be read from the position M and M * • 2. Batteries Two ignition dry cells, no, 6 IGN* 1 l/2 volts, Rayco type batteries were used In series to furnish a current which would produce a measurable fall of potential in the main circuit of the potentiometer, 3* Standard Cell An Eppley standard cell of Eppley Laboratory, Inc*, was used to regulate the current flowing through the wire AB of the potentiometer (Figure 9) and to make It exactly one-fiftieth of an ampere. The electromotive force of this cell is 1.01921*. volts at 23° C. and has a temperature coef­ ficient which Is negligible within the ordinary range of room temperatures• 39 It* Galvanometer The galvanometer used was a No* 21+20 Leeds and Northrup Portable Lamp and Scale instrument* The galvanometer’s function was to indicate the absence of current flow when connected in series with the unknown electromotive force between the regulated contact points M and M 1 of Figure 9. This galvanometer has a resistance of 1000 ohms; a period of 3 seconds; a sensitivity of 50 megohms; and is critical­ ly damped on open circuit* 5* PC Power Supply A Model 600 B Sorensen and Company, Inc., DC variable voltage power supply was used to impress direct current through the specimens in some of the static and dynamic tests. The supply was also used to apply EMF across two platinum electrodes of the electrophoresis U-tube apparatus to determine the sign of electrical charge of colloidal particles. This instrument can be adjusted easily to any output voltage from 0 to 600 volts and is of negligible ripple. ko EXPERIMENTAL PROCEDURE A, Static Tests De-ionized water used in all tests was first passed through the two-bed ion-exchanger in which the conductivity indicator indicated total solid content of the treated wa­ ter to be less than one ppm equivalent of NaOH, and then through the small mixed-bed ion-exchanger• Saturated lime solution of certified Fisher reagent Ca(0H)2 , and carbon dioxide were added to the de-ionized water to produce water of the desired levels of pH and hardness. Specimens were either cast iron or stainless steel. Cast iron specimens were 3*0 x 1.0 x 0.090 inch wafers sliced from cast iron bars and finished by surface grinding with a IgO-grit diamond dressed wheel. cimens were 3.0 x 1.0 x 0.019 inch. cleaned for five minutes Stainless steel spe­ The specimens were in 0.1N HC1, washed with boiled de-ionized water, rinsed with acetone, dried and weighed before each test. Specimens were totally immersed in make-up xvater of the desired pH and hardness, as described in the section on apparatus for static tests. The water was kept saturated k-1 with oxygen and stirred throughout each run by bubbling air continuously and by keeping the cylinders open to the at­ mosphere . In some tests a direct current was impressed through the specimens from the dc power supply by connecting the negative pole of the power supply to one specimen, which acted as a cathode, and the positive pole to other speci­ men, which acted as an anode. A 100,000-ohm variable decade resistance and model 37h- Simpson ammeter were hooked in se­ ries with the specimens. The duration of each run was exactly one week. Hard­ ness and pH were adjusted to the desired level periodically throughout each run. All tests were conducted at room tem­ perature * B. Dynamic Tests with Recirculated Water The circulating water was produced in the same manner as for the static tests, by adding saturated lime solution and carbon dioxide to demineralized water. 'Water was cir­ culated through the test cell in which the specimens were mounted as has been described under the apparatus for dyna­ mic tests . In most runs, oxygen content of the water was kept within one ppm of saturation at room temperature by using 30 gallons of water in a 55-gallon reservoir with air fil­ ling the space above the water. The reservoir was also opened to the atmosphere several times during each run. Some runs were conducted with water of different le­ vels of dissolved oxygen. Some of these runs were made at a dissolved oxygen content below that of saturation while others were at levels higher than saturation. In low oxy­ gen level studies forty gallons of water was held in the 55 gallons reservoir. Nitrogen was bubbled through this water at the beginning of the test for a period long enough to reduce the oxygen to the desired, level. Different parts of the system were then connected so that the system was air tight. Care was taken to prevent any leakage of gases Into or out of the system during the addition of lime solution and. carbon dioxide. In this way it was possible to keep the oxygen content of the make-up water within one ppm of the desired level below saturation. By following the same procedure as above and substitu­ ting oxygen for nitrogen, the dissolved oxygen content of the make-up water was maintained within one ppm of a de­ sired level higher than that of saturation. The pH and hardness of the water were adjusted peri­ odically during each test. Specimens were cleaned and weighed before each test in the same manner as for the sta­ tic tests. All runs were conducted at room temperature, k3 and at a linear* velocity of 2 feet per second. As in the static tests, in some runs a dc potential was impressed on the specimens, which acted as electrodes. The 100,000-ohm variable decade resistance and model 37^1 Simpson ammeter were connected with the specimen as shown in Figure 10. Polarization characteristics of the elec­ trodes were measured at given intervals during a run by ob­ serving the potentials of each electrode against that of saturated calomel electrode at different applied current densities. The applied current density was kept constant throughout each run except when polarization measurements were made. During polarizing measurements different values of current densities were applied for few moments only. In the majority of dynamic tests no outside current was Impressed through the specimens. Corrosion potential of each specimen was measured at several intervals during a test by comparing the potential of the specimen with that of the reference saturated calomel electrode. Change of potential values of the specimen with time was a function of the action of the inhibitor on the corrosion tendency of the specimen. The potential of she specimen at the begin­ ning of the test or at zero time was determined by immer­ sing a similar specimen, cleaned in the same manner, beaker containing the same water as that of the test. potential of the specimen was then quickly measured. In a The h1th hU. s» 0 u 0 +<» flj ? v__ . ° °13 <0 H© go r st C s s C ^ £ *a r<, 2 •£ GO U 3 &0 w nj a> a * h ^ Q W © o O ciJ aj © m this method, the potential of the specimen at zero time could be determined more accurately than when the specimen was inside the test cell. It required only a few minutes for the recirculated water to fill all lines of the system at the beginning of each test. The change of potential of the specimen during this time was very sharp. C. Specific Investigations 1. Potential Measurements Potential measurements were made using a potentiometer and standard cell technique. The wiring diagram for these studies is shown in Figures Ij_, 9* and 10. The standard cell, galvanometer, and two batteries in series were con­ nected to the main circuit of the potentiometer as shown in Figures 9 and 10. The current flowing in the wire AB was adjusted to one-fiftieth of an ampere by setting point T to correspond with the electromotive force of the standard cell (l.0192l|-). The resistance B was regulated until the galva­ nometer showed no deflection. The electromotive force of the electrochemical cell, in which the two electrodes were the specimen in the test cell and the saturated calomel electrode in the line, then replaced the electromotive force of the standard cell In the potentiometer circuit. The contact points M and M» were manipulated until the galvanometer showed no deflec­ tion. The current was checked again by switching back to the electromotive force of the standard cell. If the gal­ vanometer showed no deflection no further readjustment was needed. If a slight deflection occurred, a small readjust­ ment of R and corresponding readjustment of M 1, after re­ placement of the applied electromotive force by that of the standard cell again, was required. The reading of contact points M and M 1 then gave the potential value of the electrochemical cell composed of the specimen and the calomel electrode. Since the refe­ rence saturated electrode was assigned an arbitrary value of zero the cell potential then was that of the specimen serving as the second electrode in the electrochemical cell. 2. Preparation of Colloidal Calcium Carbonate Two procedures were followed for preparation of col­ loidal calcium carbonate: a. A supersaturated solution of CaC03 was prepared by addition of saturated lime solution and carbon di­ oxide to de-ionized water. Excess lime solution was then added to raise the pH to a value higher than 10. At this high pH colloids were formed and were prevented from growing into crystalline size by quickly adding CO2 to bring the pH to a lower k-7 value. This procedure resulted in the formation of a CaCO^ suspension with a mixture of colloidal m a ­ terial which stayed in suspension and of larger crystals which precipitated, b. Saturated lime solution and carbon dioxide were added to the demineralized water until colloids were observed at the desired pH level. that It was found by using this method, much finer colloids were produced than with the previous method. In most of the tests conducted In which the water contained colloidal CaCO^ this latter method was utilised, 3. Determination of Electrical Charge of Colloids Electrophoresis methods were used to determine the electrical charge of colloidal CaC0 3 # The solution con­ taining colloidal CaC03 was placed in a U-tube under the Influence of an EMF applied through two platinum electrodes, one in each leg of the tube. Migration of the colloids then indicated the charge on the colloids, was observed U-tube This migration by noting the turbidity In the two legs of the after a few minutes of applied E M F • In cases where the solution contained a very small amount of colloidal CaC0 3 , the migration could be seen by observing the scat­ tering effect of a light beam at each electrode. 5» Analysis of Specimen Coatings After the Tests Mineral examination was performed by the Michigan State University Geology Department by petrographic methods. Index oils of 1.58 and 1.80 were used. The mineral below 1.58 was calcite (calcium carbonate), the material above 1.80 was limonite (hydrous ferric oxide, chiefly goethite), and the Intermediate mineral was siderite (ferrous carbo­ nate) . The percentage of each mineral was estimated by the above methods and identification and percentage values were checked by use of dilute hydrochloric acid for effervescent comparison. The minerals were also studies under crossed polarizing prisms, and the properties observed corresponded to those described for these minerals In textbooks on pet­ rographic mineralogy. 5. Routine Analytical Determinations The pH was determined us in,, a Beckman Model H2 Glass Electrode pH Meter. Analyses hr total hardness and calcium hardness were by the Versenate method described in HachVer Catalog (33) of Hach Chemical Company. UniVer-1 indicator powder was used for total hardness and CalVer-11 indicator powder for calcium hardness. Alkalinity, total Iron content, and dissolved oxygen were determined by procedures described In the 10th edition of Standard Methods for the Examination of Mater, Sewage, h-9 and Industrial wastes (3^). Total iron was determined using the Phenanthroline method. The Alsterberg Modifica­ tion of the Winkler method for dissolved oxygen was fol­ lowed with the exception that a round-bottom 25>0-ml flask with rubber stopper was used to collect the sample instead of the regular B*0*D* bottle* The rubber stopper con­ tained two holes through which small glass tubes were In­ serted* It was possible, by this arrangement, to fill the flask with nitrogen gas before it was used to collect in the sample for dissolved oxygen determination* 50 CONDITIONS OP THE TESTS AND ANALYSIS OP THE COATINGS DEVELOPED Table 1 shows various conditions of static tests and provides data concerning examinations and petrographlc analyses of the coatings developed. Dynamic tests are shown in the same manner in Table 2. All tests temperature. (static and dynamic) were conducted at room Duration of all static tests was one week. Recirculation of the make-up water of all dynamic tests was at a linear velocity of 2 feet per second past the speci­ mens. In both types of tests water containing colloidal CaC0 3 , the hardness reported included that of CaC03 collo ids. In both tables, where the test numbers carry a star, CaC03 colloids were prepared following the second procedure discussed under experimental procedure for preparing col­ loidal CaCC>3 . That Is, lime solution and C02 were added to the demineralized water until colloids were observed at the de s ire d pH leve1. In Table 1, an analysis of the coating of only one specimen is shown for some tests although two specimens were used In all static studies. The coatings of the other corresponding specimens were removed for weight loss 51 determinations. The test cell which was designed for only one specimen was used in some dynamic tests where the ana­ lysis of the coatings of only one specimen is reported. Langelier!s chart (6) was used to determine "saturation indices” graphically. "Stability indices” were cal­ culated on the basis of pHs values obtained from Langelierfs chart. Values of "saturation excess” were deter­ mined experimentally using the "marble” test. The equation developed by Dye (32) was used to calculate "momentary ex­ cess” levels. Terms used to describe coatings are as follows: Blotchy: Irregular spots of large concentrations. Powdery: Pine, loose particle appearance. Tarnished: Discolored dull appearance. Sugary: Granular appearance resembling sugar. Spotty: irregular spots of small concentrations. Sparse: A coating of negligible amount. Platy: Tendency to break off into sheets. Table 1 Static Tests Anal Conditions of the Tests ^ i ir>1,3 Total .hardn e ss ppm e cjuiv. of jcaco^ ;o 12 r -at ^■ .r-.-O-'OO<5^6 d 7■' , Onrq / *r^ :\c J ' ' 11 i 92 cast iron oa tiiode 1.03k Cast iron anode Iron c n 1'Q lo j.jz ir~n m ;de 92 8.6 Cast iron cathode Cast iron anode Cast iron specimen 02 Other spec ifications of make-up ■water Satu- Sts biration lily index index I men* t,cry ereon' Cal- .0 oid** 3 Si­ s-ram cite erite mon It e s ize CaCOg PeCOy FeO(OK) (mm) ,0 In addition, 1 grn of ground CaC03 powder was added. + .65 92 88 20 70 15 80 + .35 + .65 7*3 7.3 IIl.3 2.7 15 20 65 l/90 + or - Pine gra i»ns, under y , Aggregates up to 1 mm. In size, fine coating between aggregates. 5 20 75 l/90 + or - Blotchy, fine gra Ins. Ik 10 15 75 l/90 + or - Blotchy, fine grains. 5 10 85 predominately between blotches, thin coating. ^Aggregates dark In center, light around periphery, calcite found l/l80 to 1/90 Blotchy, fine grains• predominately between blotches. 8.6 Cast iron specimen 1 92 Contained col­ loidal CaCOg, 65 •J 92 Powdered CaCC>3 + .65 ground in a ball mill for 6 da:ys was used to form the make-up water 7.3 + .15 8.3 Cast iron specimen 2 Cast iron specimen 1 8.6 50 Pine grains, he avv coating of ru very soft. l/l80 to 1/90 Pine gra?_ns, heav; coa t ing c f r■ .\st , very soft, Pine grains, aggregates distribution, uniform coating, 0.25 mm. in thickness, 1/180 to 1/90 Blotchy, -U '■* 1LS , Scrapes off in plates, fine grained aggregates, uneven coating. slightly nar han anode co-kin * l/90 + or Blotchy, n no 13 . Pine grained aggregates, uneven coating, coating heavier in cent< f ' of plate. j. 1 br.p 2.7 10 15 75 l/l80 to 1/90 Pine gra ins, 7neven coating • •Aggregates separated by fine coating, lk. 5 2.7 10 15 75 l/l80 + or - Powdery, ['Aggregates of fine grains. Ik. j-.u 2.7 10 10 80 l/90 + or - Blotchy, fine graIns. 10 15 75 l/90 + or - Blotchy, fine grains. 5 20 75 l/l80 to 1/90 Pine grains, uneven coating, soft. 1 Cast iron specimen 1 Pine grains, heavy coating easily rubs off, coating l/8 mm. in thickness. 10 l/90 to l/lk • 7*8 Contained col­ loidal CaCOy, Micro description 85 10 7*3 Macro descrlot ion 10 71s of ground CaC03 powder on 1y was a dde d to form the 111ak e -u p wa t er . 031 1.032 Satu­ ration excess ysis of the Coatings + O.k Jalcite predominates between blotches, thin film of calcite over blotches, rubs off readily. Pine grains, uneven coating, ridge type, few distinct crystals* mostly aggregates. Table 1 (II) Static Tests n a l )jysis Conditions of the Test; 01 the Coatings Total t \ -> u )TI *n ,iiv . o1 CaCQg d.b 60 9.0 92 Cu ” ''i t liin 7 °sed 0 Contained col loidal CaCOy. 19 1 1 10 1 iron cations of make-up water Satu- eta; i» ration 1 t ind« ^ T.Q. Saturation exce; Io 1 1 t^ * c •< 3 Cal- -/o Sid- /0 LiGrain cite er it e raon ite size CaC 03 FeCQp FeO(Oh) (mm) + .15 8.3 O.li 5 20 +1 •Op b •° 11.1 20 25 55 1/90 to 1/25 25 60 1/90 to 1/25 10 85 1/90 + or - ‘'i't' i C "” *r i'i oc m~r> 11 Chs jn ~ "Oi 1 9.0 91 12 Cast iron specimen 1 9.0 50 Contained col­ loidal C aC 0 0 , +1*05 6*'' + .56 7.' 11.1 75 1/90 + or Macro de scr ipt ion Micro description Blotchy, fine grains. Fine grained aggregate, uneven blotchy coating. Blotchy, fine grains. ;Coating readily scrapes off, uneven coating, well formed calcite crystals scattered throughout, Blotchy, fine grains# pust aggregates through specimen, well formed calcite crystals found close to plate between heavy concentrations of liraonite. Blotchy, darker staining in center of blotches, 1Calcite crystals formed between blotches of siderite and limonite. 1 B1 ot ch y , f in e gr a ins , iFIne grained blotches of siderite, limonite, and calcite, well rn.bs off. 0 3.3 ily. formed calcite crystals "between blotches , j Contained col­ loidal CaC-0 3 . 10.23 13* Cast iron specimen 1 +2.20 5 . 0 66 .P U.5 Cast iron specimen 2 Ik 1$ 16 Plat in UTl cathoc e Pla t inum anode Stainless steel cathode Stainless steel anode Stainless steel cathode 92 9.3 .032. 7.3 lk*5 *i 15 1-5 1-5 95 1/90 + or 1-5 1-5 95 1/90 + or 1/90 + or - 100 1/90 to 1/1|3 100 l/90 + or - 2 gms of ground CaC03 powder on 1 j was added to form the make-up water. 13 92 + .65 10 1.03k In addition, 1 gm of ground iaC03 powder was added. + .65 7 •J Ik. r 100 l/25 + or V ery uns ven clis tr ibution, rubs off easi­ ly. Very uneven distribu­ tion, soft. Large uneven blotches of limonite, very soft, readily crumbles ; off. Massive uneven concentrations of liraonite, calcite found through; out next to plate. of crystals. Fine grains, even feven distribution, voids 1 to k time. coating, Crystals present at one end of plate, voids 1 to 7 time3 size of Fine grains, uneven distribution located crystals. at end. Tarnished Very fine film. Fine grains, even distribution. Fine crystals, concentration poor, voids 1 to 10 times size of crystals. Table 1 (III) Static Tests Anal'ysis of the Coatings Conditions of the Tests Total hard­ ness PP1:1. eaniv. of CaCO ■3- Ourrent impressed ma/d:m2 Other spec ifications of make-up wafer Satu- Sta bi­ Satu­ ration nation lity excess index index Momen­ tary excess % Gal- % Sid- % LiGrain cite erite monite size CaCO3 FeCO3 FeO(OH) (mm)__________ Macro descrip tion Micro d e s c r i p t i o n _____________________ Less than 1/180 Very fine film. Very fine grains, presence of calcite indicated by violent reacI tion with acid. 100 1/90 to 1/Ll5 Fine grains, even coating. .Fine grains, even coating, voids few, those present 1 to 2 times size of crystals.- 100 Less than 1/180 Very fine grains, ’Very fine grains, uneven distribution. 100 l/90 to l/lj.5 Fine grains, uneven coating, Very few crystals or aggregates, coating present only at the cen­ ter and at ends of plate. Tarnished. .Very fine film, no reaction with HC1 . Fine grains, even coating. '(Spotty crystal distribution, voids 1 to 5 times size of crystals < 100 hh:t inl e s s 3 13 e1 _ --- an O',is Ttahnlo3o Sf y o 1 0at'".ode 2fa.in less st 3 el arc to 1.0 3k 3 ’ i* ilo 3 "tool ^1t o h ■\'il "0 31ee 1 1 ' <3 92 o ta iniess stool cathode 3to inlens st oe1 a?.10 do 3.6 Stainless steel cathode 3tainless steel anode o *o Stainless steel cathode Stainless steel anode + *35 + 3. Oh 3 2 *06 3 C on ta in ed col­ loidal CaCO^, 0.517 92 0.517 55 + .65 + .65 Contained col­ loidal CaCOg. + .65 1I4-.5 7*3 1/. 5 7. ill.5 7.3 Ik.5 7*3 lit.5 7.3 1 2.7 2.7 l/90 to l/i[-5 2.7 2.7 2.7 Tarnished, Very fine film, no reaction with HC1, 100 1/90 ...... Few crystals scattered widely on plate. 100 l/90 + or - Tarnished. One or two crystals per 100 1/90 to 1/30 Fine sugary coating, Few calcite crystals and aggregates, concentration poor. ------- Tarnished, jNo reaction with HG1, inch square area. Table 1 (IV) Static Tests ondit ions of tie Tests Anal ysis of the Coatings T otal hand119 S 3 pan equ. iv of ■13 22*'c 2 b a I-7.1ess 9.1 100 Cnrrent Impreased "/in •na nm*1,031' 3 t c o .1. Other spec I:ficat sens of mahe-up rater Contained col­ loidal OaCO-/, Catti- 3 tab Iration lity Index index +1.25 Saturation excess Momentary excess 10.2 6.6 cathode D ta .1113.33 3 s t e 3*1 5 Cal- % Sid- % LIGrain cite erite raonite size CaCOy FeCOp FeQ(OH) (mm) Micro description Macro description 100 l/90 to l/ij.5 Spotty coating Voids 1 to 5 times crystal size. 100 1/90 + or - Spotty coating Voids 1 to 5 times crystal size. 100 1/90 + or - Granular coating, Amorphous calcite, 1/33 crystalline granular appearance, voids 1 to )_l times crystal size. 1000 1/90 to i/k5 spars e dis tribut ion• Crystalline and amorphous calcite present. 100 l/90 to l/25 Sparse coating, Coating uneven, 100 1/90 + or - Uneven coating. Coating uneven. 100 l/90 + or Fine grains, 'uneven distribution. One side of plate was even grain size, the other side showed varl ation in size l/90 to l/l80. 100 1/90 Fine grains, uneven distribution. Even size grains throughout, voids 1 to 10 times crystal size. 100 l/90 and 1/180 Uneven distribution. sEven distribution, voids 0 to 10 times crystal size. 100 1/90 + or - Uneven distribution. Uneven distribution, size even, voids 0 to 20 times crystal size* 100 1/90 + or - Even coating. Voids 1 to 5 times size of crystals. a n 3 ri.fr 23* Stainless steel CatliOdG 8tain1es3 3tG61 :ii'DUO 0 2Lj* Stainless steel cathode Stainles s steel anolo 70 25* Stainless s fceel cathode Stainless ste eJ. an ode 9 el 100 26*“* Stainless s teel cathode Stainless steel anode 27* Stainless steel cathode 130 10. 83 1.03i|- Gontained co 1 • loldal CaOOp, Contained col­ loidal CaCC'3„ 0'}h 035 Contained col­ loidal CaCOy. Contained col­ loidal CaCO-^. 1 . 0 C on■ ta ined col­ loidal CaCOy. Ik. 8 + 1.70 +1.35 +1.60 +1.65 +2.20 lk. 9 6.a 17.1 5„9 22.1 6.2 5.85 66.5 k6.5 Table 1 (V) Static Tests Anal ysis of the Coatings Conditions of the Tests ^ "C i_ )> ^ J b-3>ul .-j+•fta ] JL10 1 { !-•ua.-l_11 fc33 Total hard­ ness jypiii equ iv _ .r» O i. CaCOp 8.8 220 «?f’ .f3.S' ] ih-: yJu i x rpec *rt c >t *^ ^ mu ”c »'1 ja be G on ta in ed ao1 loidal CaCOo. Satu- Stahl* ration lity index index +1.60 Satu­ ration excess Momen­ tary excess 13*.5 5,80 3 e ^ * i„ i 1 3 jn i ^L »"b "t i' J " l ~\ Z ) 11" t t 3..*.31 '•3~ st eel 3 c>*■> 1 u3:ii ^ L \ 2 90 Contained col­ loidal CaC03 and iron oxlds +1.S5 2k. 0 6.30 | Grain ;i Cal- ;b Sid- % Lic ite erite monite size CaCOg PeCOy FeQ(QH) (mm) ricro description Macro description 100 l/l80 to l/90 Sparse distribution. Coating easily rubs off, voids 1 to 10 times crystal size. 100 l/90 + or - Fine grains, uneven distribution, Crystals and aggregates tend to build up on one another, plate ’is pitted, crystals located around, it. 100 l/90 + or - Fine grains, uneven distribution, Coating good, few voids, crystals closely packed. 98-99 1/90 + or - Fine grains, hard coating. Calcite discoloration due to Fe0(0H) present 1 - 2%» 98—99 l/90 + or - ^ine grains, hard coating. jbarne as Spec imen 1 I t ! i | Table 2 Dynamic Tes r ■fn a l y s ijisort h e Coatings Ton Titians of the Tests Test To . u2hH.2':-un3 Total hard­ ness ppm equiv. of jiaCOs, Oast iron oafRode Casa iron anode 92 Cast iron cathode Cant iron anode 92 Cant iron cat ho deGast iron anode Cast iron specimen I Current impress V.ia/-+ Durat ion of t O3t (daps fis s olved oxason (ppe . Other spec if icat ions made-up water Saturat ion + «6p + ,61 Satu­ ration 3.3 ISatu£ tab 5. ration lity index in dee: t .35 O 9. 7.3 .3 7.3 12*5 lij.3 Ik ,5 erite FeCQ^ 13 15 70 10 10 80 pp 25 1+0 1/90 + 20 15 65 1/90 + o:- - k.0 20 1+0 - - Platy coating. Peels off in plates or sheets, 30 20 50 i/30 F in e grains, 1ine ar coating. Linear arrangement, calcite predominant on top of coating. 20 15 65 1/90 ■* 19 10 75 1/90 + or- Linear spotty coating, 10 10 80 l/l 90 Hard linear coating. 20 10 70 2.7 10 70 k 6 .5 30 30 2.7 2,7 2.7 7“ Cast iron 10.25 soec iraen 1 Grain s ize (mm) .'iacro description r - ration O9 Saturat ion 92 3 a tu r a t ion Cast iron specimen 2 Cast iron specimen 1 * % Da­ mon it e FeO(OH) ; + .65 7.3 Ik. 5 2,7 Cast iron specimen 2 Cast iron specimen 1 % Sid­ far c if e GaCQg SatuMo menration tary excess excess C on ta in ed co 11 o ida 1 CaC03. 92 Saturation Contained colloidal CaC03 ground in ball mill for 6 days. 83 Saturation Contained colloidal CaC O3 • 3 |+ op + 2.20 (• 3 Ik •9 l.k, 00 ,p 2,7 r - 1/25 or - 1/90 + o- Micro description \Fine grains, linear coating, Fine grains, linear coating. Linear arrangement, calcite was predomi­ nate 011 top of the coating. Irregular linear coating. F ine gra Ins, 1 inear coating, Fine grains, linear \ coating. Linear arrangement. tLinear spotty coating, 1Hoarse hard coating, 1 in ear a r r an geme n.t . i iieavy coat ing, calcite ■rubs off easily. *Linear arrangement, soft and loose. Ridges of limonite and siderite, calcite in valley3 , amorphous calcite prominent over coating. Ridges of siderite and limonite, calcite present overall. Calcite readily scrapes off. Similar to Specimen 1, scrapes off more easily. Particles vary in. size from microscopic to aggregates, limonite in linear ar­ rangement to calcite. Linear arrangement, hard coating, glassy texture on Te mineral. Fine powdery calcite aggregates overlying plate, ridges of limonite and siderite*. Calcite concentration in linear arrange­ ment, also siderite and limonite linear concentration. Table 2 (II) Dynamic Tests -nalysis hof the Coatings Conditions of the Tests Total hardness ppm equiv. of GaCOg onoe .mens Dura­ tion Currm’t of lire "S3cd test 1 '/ 1 (days) DIssolved oxygen (ppm) Other spec If i« cations make -up water5 Saturation Index Stalllity Index Satu­ ration excess Momen­ tary excess 45 :isty iron spec iraen 2 ist iron 10 .20 specimen I ^ i j- i pr g i t ’r ~ » c■1me 2 o ,o J ]f iron ■'cI 1m. '1 ,Q >-> IV' Cast iron specimen 1 150 Satu­ ration Contained colloidal CaC03 „ +2 •0 2 6.15 48.5 3)+.7 Satu­ ration ContaIned colloidal CaC 0 3 . +iao 6.4 30.0 5.1+ 12v Cast iron specimen 1 Ctainiess steel specimen 2 sand blast 3„3 13" Cast Iron specimen 1 3 % Sid­ erite FeC O3 25 1-3 monite Fe0(0H) Grain size (mm) - Macro description . — *km ■■ ^ ,.i »i Micro description Linear, calcite loose. Fine grained aggregates, linear arrange* ment, as in Specimen 1 . 30 1/90 + or 95 1/180 to 1/25 Irregular coating. Linear arrangement, light calcite film between limonite concentrations. Me d ium ha r d , 1 inear coating. Medium har d , linear coating. Arrangement similar to specimens of Test u Linear coating, bonding not as good as Specimen 1 . 20 15 65 l/90 + cm - 35 20 45 l/90 + or - 120 Saturat ion Contained colloidal CaCOy • +■ 1.0 6.7 22.0 5.1 10 10 80 1/180 + or Medium hard, fine li­ near coating, Fine linear coating, granular calcite overlying plate, ridges of limonite and siderite. f 203 Saturat ion Contained colloidal CaCOo „ +1 *15 6.1 44.0 5.0 30 10 60 1/80 + Medium hard, fine coating. 30 10 60 1/100 + or Medium hard, fine coating. Calcite overlies limonite and siderite, limonite found next to plate, fine gran-' ular coating. ^ Coating similar to Specimen 1. 25 10 65 1/100 + or ’Fine, hard coating, Cast iron speciraen 2 sand blast Stainless steel specimen 2 % Calcite CaCO^ 200 200 Satu­ ration Contained colloidal CaCO^,. + 1.05 Satu­ ration Contained colloidal CaC03 . + 1.05 6.2 44.0 i+.o hen coating, grained calcite coating overlying limonite and siderite. ho coating visible. 6.2 44«o l+.o 10 60 1/180 + or Fine, hard, sugary coating• ]Mo significant deposit. Even coating tending for linear arrange­ ment , calcite overlies siderite and li­ monite . Table 2 (III) Dvnamic Tests malys is\ of the Coatinvs Conditions of the Tests Total hard­ ness ppm equiv« of CaCOp ■3pec imon: 200 Cast iron spscinen 1 Dura­ tion of Current impressed test (days) ma/dh- 0 1 Dis­ solved oxygen (ppm) 0th or spec If Icatlons r make-up water ,3aturation Contained colloidal CaCO -j« 6.2 1(4.0 4*0 40 10 50 l/lS o + j? - ¥ 4.0 15 5 75 1/90 ivO 1/180 • :Hard coating, linear j arrangement. hi! 4.0 10 10 80 1/90 to l/l-O Soft coating, linear arrangement• Ridges of limonite flat and tend to break off into plates, calcite mainly in the valleys. 307 10 5 20 l/l80 -1- or Une ven, medIum hard coating, Contained 03.j magnetite, linear arrange­ ment, calcite found in between ridges of ma.on e 111 e an d siderite. 30 20 50 1/180 to l/90 Soft coating, linear arrangement. riderIte and limonite predominate close to plate, amorphous calcite overlies coating and predominates in valleys. 20 15 63 1/90 Even, soft coating. 20 15 63 1/90 Even, soft coating. Contained 2R magnetite, limonite harder, siderite and magnetite coating to plate overlies by amorphous calcite. Contained 2/ magnetite, similar to Speci­ men 1 , however calcite more intermixed instead of predominating over coating. 10 10 80 1/180 Fine grains, soft coating. Saturation 0 onta ined col- tl.O1 loidal Oat03. Cast iron specimen 1 3.3 200 Saturati,on Contained col +1«05 loi da 1 Ca'0 g. 1*5 2.5 Contained col­ +2.05 loidal CaiOp. 200 1.7 2.7 Contained col' +1*05 loidal Ca2'03. 200 27 28 Contained col*i+l*>05 loidal Ca.■Op. 0.3 v Cast iron specimen 1 6.2 .13 .8.3 !J J ± ( O. u-. c 48 k.O Cast iron spec iraen 2 ■* Cast, iron specimen 1 0.3 0.88 2.1 Grain s ize (mm) +1.05 200 Oast iron specimen 1 % Li­ monite FeO(OH) lity index 8,3 Cast iron 10,2 specimen 1 / Sid­ erite FeCOy ration Index ' Cast iron 3nocinen 1 Momentary excess % Cal' cite CaCOy Saturation excess Contained col- + *95 loidal CsC03. 6 •4 3.2 Macro description Micro description Fine grains, soft coating. Soft coating. Contained 55 magnetite, liraonite ranges in hardness from 2 to 5, blotches of hard limonite, calcite overlies soft limonite, magnetite and siderite dis­ persed throughout. Contained 11 magnetite, coating flaky tends to break off in sheets, very lit­ tle calcite next to plate, limonite, sid­ erite and magnetite predominate next to plate. Table 2 (IV) Dynamic Tests Analysis I Conditions of the Tests e icis dH Total hard­ ness ppm eauiv< of CaCGp a31 Iron specimen 1 8. Ida :ast iron specimen 1 3,3 1’ aot iron specimen I ■ast iron specimen 2 3, j 130 last iron specimen 1 .3 130 Dura Dis­ t Ion solved of Current oxygen impressed test (days) ma/ dm? 0 Other specifi­ cations of make-up water Satu: ,ration iindex Stability Index Saturation excess liomentary excess 100 % Sid­ erite FeCOy % Li­ monite FeO(OH) Grain s ize (mm) Macro description Micro descrIption Contained col-j-t ,95 loidal CaCOp* ■ 6 .k 3.2 25 15 60 1/180 Fine grains, soft coating * Contained 1,1 magnetite, linear arrange­ ment, calcite overlies ridges of limonlte and siderite, calcite also found in valleys, intermixture of all present near plate. Saturation Contained col-|+ .95 loidal CaCOp. { 6 .l± 3*2 25 20 55 1/180 Linear arrangement, medium hard coating. Contained. 1/ magnetite, blotches in li­ near arrangement, amorphous calcite intermixe d thr oughout, Saturation Contained col- + ,95 loidal CaOOg# 6 3*2 25 20 55 1/180 25 20 55 1/180 Linear arrangement, medium hard coating. Linear arrangement, medium hard coating. Limonite predominates near plate, .cal­ cite overlies coating. Same as Specimen 1 . 10 10 78 1/90 Uneven, medium hard coating. 10 10 79 1/90 to 1/180 Uneven, medium hard coating. Contained 2,5 magnetite, irregular blotchy coating, blotches in linear arrangement, calcite overlies plate, limonite, sid­ erite and magnetite next to plate. Similar to Specimen 1. 1-2 1/180 Uneven, medium hard coating. 3 10 11 Contained col + *95 loidal - m il 6 .JLl 3*2 'ast iron specimen 2 tainlets steal specimen 1 % Cal­ cite CaCOg Coatings Saturat ion Container c c h +1*05 loidal CaOO^ and iron oxide. 5,2 kb 98 Amorphous calcite coating more prominent where limonite present, coating tends to be at one end of plate and opposite on g the other side, voids 1 to 10 times size of particles. 61 DISCUSSION OP RESULTS Mineral Content of Developed Coatings Examination of minerals developed upon the cast iron specimens showed that coatings consisted of mixtures of li­ monite (hydrous ferric oxide, chiefly goethite), calcite (calcium carbonate) and siderite (ferrous carbonate). Mag­ netite was also present in some tests, usually covered by limonite. With stainless steel specimens only calcite was found. Microscopic examination revealed a linear arrangement of ridges and valleys on cast iron specimens from the majo­ rity of tests. This linear arrangement was particularly noticeable for specimens from dynamic tests. Most of the calcite obse rved was found In the valleys, while ridges were commonly limonite on the surface with siderite near the metal. Calcite showed typical rhombohedrals on the stainless steel specimens. In the valleys of cast Iron specimens some calcium carbonate crystals were noted but the calcite was commonly amorphous. Ridges were predominantly hydrous ferric oxide in the form of limonite with some calcite of­ ten present on the top. Siderite was commonly found In very fine crystals. Limonite was not crystalline, being observed in the amorphous form only. Percentages of the siderite constituent of coatings very closely followed that of calcite. Magnetite was found in amorphous form only. Polarization Studies The significance of polarization studies In Investiga­ ting the action of a corrosion inhibitor has been discussed previously under theoretical considerations. These studies were undertaken for the purpose of determining the function of calcium carbonate as a corrosion inhibitor. In the first three dynamic rims, studies were made using a galvanic couple of cast iron specimens mounted in the test cell of the flow system. These runs were conduc­ ted at a pH level of 8.6, a hardness of 86 ppm equivalent of CaCOy, water saturated with dissolved oxygen, and no colloidal CaCOy present. The applied current density was the only variable in the testing conditions of the runs. Direct current was applied through the specimens, which acted as electrodes, from the dc power supply. Cur­ rent densities of 0.826, 1.652, and 3-30if ma/dm.2 were re­ spectively applied in these three runs. Polarization characteristics of the electrodes were determined according 63 to the method discussed under experimental procedure* Figures 11, 12, and 13 show polarization characteris­ tics of the anodes and cathodes for the three tests* In each case the upper curves represent the potential values of the cathode at the Indicated rate of applied current density between the electrodes. The lower curves show the potential-current density characteristics of the anode. Percent compositions of the coatings developed after 7 days period are shown in Table 3* Figure 11 shows polarization curves which were ob­ tained at the applied current density of 0.823 ma/dm^. These curves show that the corrosion was essentially under mixed control for the entire test. At applied current densities of 1.652 and 3*30li ma/dm^ Table 3 indicates that more calcium carbonate was deposited on both cathodes and anodes. Polarization curves obtained at these higher current densities are shown in Figures 12 and 13* Comparison of the curves in Figures 12 and 13 with those in Figure 11 reveals that the only noticeable change which took place as result of deposition of more CaCOg was increased polarization of the cathodes as the tests pro­ gressed. This increased cathode polarization, as measured by the slopes of the cathode curves, was more pronounced in Figure 13 where more CaCC>3 was deposited on the electrodes. Open-circuit potentials of the cat/code curves of Figures 12 ■M. Polarization Curves for Dynamic Test 1 - 2.6- -2.Z- C urrent F i g u r e 11« D e n s ity = . 826 m a /d m / C a s t Iron S p e c im e n s D y n a m ic T e s t ® 0 2 hours 24 hours d 50 hours 0 89 hours A 100.25 hours -Z o - 100. 25 h r s . 89 h r s . 24 h r s p o te n tia l--v o lts cathode Electrode 2 hrs. 2 hrs. 24 h r s . potential of saturated calomel electrode 89 h r s . 100.25 hrs. *2 - 5 /O /5 Applied Current Density m a / d m ^ £o 50 h r s . anode ^5 Polarization Curves for Dynamic Test 2 8 6 F ig u r e 1 2 . C u rr en t D e n s ity = 1. 6 5 2 m a /d m ^ 136 hrs . Cast Iron Specimen D ynamic Test 109. 5 hrs 4 2 85. 5 hrs 60.5 hrs 0 8 cathode 6 hrs. 4 2 0 81 hours hours .5 hours .5 hours 109. 5 hours 136 hours 6 4 2 hrs 0 2 4- 60 anode 136 hrs. 109.5 hrs. 6 0 ---- rt---- '---- 1*5---- '---- 2?0---- 1 ---- zs ^ ---- 1 Applied Current Density m a / d m ‘ 66 Polarization Curves for Dynamic Test 3 F ig u r e 1 3 . C u rr en t D e n s ity = 3. 304 m a /d m 2 143 hrs. Cast Iron Specimen D y namic Test 101 hrs. 52 hrs. Cathode Electrode potential--volts 22 hrs. 6 hrs. 0 ,6 Os o potential of saturated calomel electrode © 6 hrs. O 22 h r s . 6 hrs. anode 22 hrs. | 52 h r s . ^ 101 hrs. 52 hrs. 143 hrs. 101 hrs. 143 hrs. /o '£ AnnlipH f.nrrpnt Qpnsitv m a ZdflCL2 $7 id o o © Ph CO o O 1A -d‘ vO IA r-1 O iH 1A 1A O i —I 1A o o -d- 1A o' O o A] VA i—1 O CM O CM Table 3 of Specimen Coatings Three Dynamic Tests for © All O O cd o Composition the First © ■d o A p cd o ■d A © a -pcm •H © *H ^ i —( fn co *d A, 3 has been deposited on the anodes. This indicated that calcium carbonate had no inhibitory action on the anodes of the galvanic couples. From the above discussion, it may be stated that cal­ cium carbonate acted primarily as a cathodic inhibitor on cast iron specimens. Table 3 indicates more CaC03 deposited on cathode than on anode specimens. The neutralization of hydrogen ions in the solution at the cathode resulted in the production of alkaline condition in the vicinity of this electrode, as has been discussed under theoretical considerations. An increase in the pH value of the solution near the cathode would then be produced which shifted the CaC03 equilibrium in such a way to favor more deposition in that region. Formation of Colloidal CaC03 It is to be recalled that two procedures were followed for preparation of colloidal CaC03 by the addition of lime solution and carbon dioxide to demineralized water. In the 69 first method, the pH was allowed to rise to a value higher than 10 by adding excess lime water to a supersaturated solution of CaC0 3 * Colloids formed were prevented from growing into crystalline size by quickly adding CC>2 to bring the pH to a lower desired level. In the second and most used method, the pH was held at a predetermined level while lime solution and carbon dioxide were added to de­ ionized water until colloidal CaCOg appeared. An investigation was made of the degree of supersatu­ ration of CaCOy solution at which colloidal particles were formed at different levels of pH, Carbon dioxide and satu­ rated lime solution were added simultaneously to two liters of de-Ionized water in a Lj.-liter flask. The addition of these two chemicals, after saturation conditions had been reached, was carefully regulated so that the pH level of the solution was maintained at a desired level. Ten mi­ nutes after each 2 -ml addition of saturated lime water and corresponding addition of C02, the solution was examined for colloidal particles of CaCC^. The scattering effect of a light beam was used as an Indication of the presence of colloids, A slight drop in pH value was another Indication of colloid formation. Table I4. shows pH and hardness values corresponding to the supersaturation levels at which colloidal CaC03 ap­ peared. Values of "momentary excess11 and "saturation 70 ormation Table Lj. of Colloidal CaC03 Particles l>5 £h cd -P £ 0 0 £ o & « S 0 [ii £ O •rl -P w 0 2 0 -P o c5 0Q © U -p d 0 <—1 cd > 0] »H C"i 0 ^ 0 0 0*0 0 cd 'd o 5h S cd f t P K ftO Isr-J CO • O' OJ O' I—I CO -d* • O' • "LA LA * vO 00 -AT "LA * vO O' O' LA i— 1 OC\J r~1 i— i 00 vO CA CO CO m CO 00 • O' vD • O' LA AJ • O iH O' vO 71 excess1’ corresponding to these levels of supersaturation are also shown. Values of pH were plotted against values of ’’momentary excess” and "saturation excess” as shown in Figure lip. This figure indicates the ’’momentary excess” and ’’saturation ex­ cess” values for supersaturated solutions at which CaCOg colloids formed at different pH levels. Values of ’’momen­ tary excess” required to precipitate colloidal particles out of solution were higher as the pH levels increased. Values of ’’saturation excess,” however, decreased as pH le­ vels acquired higher values. Condensation of molecules and ions into colloidal par­ ticles has been examined comprehensively by von Weimarn (3 5 )- He concluded that relative supersaturation, not'ab­ solute supersaturation, governed the rate of nucleus forma­ tion of these particles. The ratio of ’’momentary excess” of supersaturated solution to the solubility of CaC03 in solution can be defined, in this case, as the relative su­ persaturation. The solubility of CaCO^ in water is con­ stant at all pH levels. Formation of colloidal CaCOg, thus, should not be influenced by the pH level of the solution and should be a function of ’’momentary excess” values of solution only. Variation of "momentary excess” values with pH levels may be due to the temporary production of local areas of 72 Momentary Excess o 1A O O ro O OJ o rH O O O P< Figure O Ilf. - Formation of Colloidal CaCO rO CO o o o o Saturation Excess o 73 very high supersaturation. Two supersaturated solutions of the same '’momentary excess" values but at different pH le­ vels contain different concentrations of calcium, carbonate, bicarbonate, and hydroxide ions. The solution of the lower pH level contains relatively higher concentrations of cal­ cium and bicarbonate ions and relatively lower concentra­ tions of carbonate and hydroxide ions. By adding 2 ml of lime water to low pH solutions, a higher temporary "momen­ tary excess" value is created locally in the areas where lime water is only partly mixed with the main body of the solutions, and bicarbonate ions are converted into carbonate ions* In addition, a heavy ooncon bration of calcium ions also exists in local areas of this solution. Colloids may thus be formed locally in these overconcentrated areas. Be- dissolving of colloidal particles once formed is a very slow process. The colloids formed m a y also act as nuclei to fur­ ther aid CaC03 precipitation. There is less chance of forming these temporary local areas of very high supers aturation by addition of 2 ml of lime water to ^Ign pH solution. There are relatively few bicarbonate ions to be converted into carbonate ions and the concentration of calcium ions is relatively low too. Calcium, carbonate colloids were also prepared by grin­ ding CaC03 powder In a ball ■‘1 for b days. Some CaCC>3 colloids were formed when this powder was added to de-ionised water. Colloids prepared in this manner were used in Dyna­ mic Test 7 and Static Test 20. It was found that these colloids were very unstable, precipitated heavily, and did not produce favorable coatings. Electrical Charge of Colloidal CaC03 It was found that colloidal calcium carbonate parti­ cles, in the pH range of 6 - 11, always migrated to the cathode when placed in an electric field of an electropho­ resis apparatus. This observation indicated that these colloidal particles carried positive charges. Such finding was not in agreement with that of Larson and Buswell (3b) who have stated txhat calcium carbonate had a negative charge In the presence of calcium bicarbonate as well as in the presence of calcium hydroxide. In order to check further, the following tests were made; 1, De-Ionized water was distilled twice and used to dissolve PIsher certified reagent Ca(0H)2 to form colloidal CaC0 3 . The resulting colloids showed a positive charge In the pH range of 6 - 11. 2. Colloidal CaCC>3 was prepared from lime water, and excess Na2C03 and C02 added. showed a positive charge. The colloid formed 75 3* Moderate concentrations of CaCl2 and Na20C>3 were mixed to prepare colloidal CaC0 3 , which showed a positive charge, 1+. CaC03 powder was ground in a ball mill for 6 days. Some CaCC>3 colloids were formed when this ground powder was added to de-ionized water. The colloids once again showed a positive charge. 5. A current was impressed through two stainless steel specimens, one acting as the cathode and the other as the anode. CaGOy. The solution contained colloidal More CaC03 was deposited upon the cathode than the anode. Upon the basis of these experiments, it was concluded that colloidal CaCC>3 has a positive charge in the pH range 6 - 11 . The electrical charge carried by colloidal particles is of fundamental importance, because without this charge col­ loidal systems are unstable. The source of the charge is explained by the so-called electrical double layer theory. This theory assumes that a layer consisting of two parts surrounds the colloidal particle. The first layer approxi­ mately a single ion in thickness, remains firmly attached to the wall or surface of the solid phase and determines the charge. The second part of the double layer extends some distances into the liquid phase and is considered movable. 76 Since a colloidal CaC03 particle travels to the cathode when placed in an electric field it may be assumed that such a particle has preferentially adsorbed calcium ions from the solution, thus localizing a number of positive charges in the immediate neighborhood of the particle!s surface. The hydroxyl ions may be assumed as the second part of the double layer which is associated with adsorbed calcium ions and is dispersed around the particle at a certain average distance from it as shown in Figure l5« The concentration of hydroxyl ions is only slightly greater around the parti­ cle than in the main body of solution. Colloidal calcium carbonate particle may thus be represented by (CaCC>3 )Ca+ + fOH“ where (CaCC^) implies a particle of calcium carbonate and the dotted lime indicates approximately the limit of the fixed part of the double layer. Effect of Colloidal CaCO3 on Coatings Developed Most static and dynamic tests were conducted with cal­ cium carbonate present in colloidal form. Hard, dense, protective coatings, with good bonding to the metal surface, were generally obtained in the presence of the colloids. This was especially true for coatings on specimens obtained with dynamic tests. Coatings from tests made under identi­ cal conditions without colloidal material present were 77 Ca (OH) Figure 15> - Schematic Diagram of Colloidal CaCO^ Particle 78 softer and formed a poorer bond to the metal, Microscopic examination and petrographic analysis of the coatings revealed that the action of colloidal CaCOj was to improve the bonding and hardness of corrosion pro­ ducts, especially limonite. Linonite sresent in coatings obtained in presence of colloidal GaGO^ was harder in Mohfs hardness scale and darker in color as compared with that formed under similar conditions without colloids. Calcite and limonite were found in physical but not chemical mix­ ture * Coatings obtained from some specimens under dynamic tests in the presence of CaCO^ colloids consisted of a uni­ form clinging protective mixture of limonite, siderite, and calcite. This material formed the entire coating including the innermost layer next to the surface of metal. This mixture bonded well to the cast iron specimens and was hard and relatively tough. Coatings obtained from tests where colloidal material was absent were quite different. In this case the layer at the metal surface was found to be formed o° loose grains of li">onifo over which layers of mixed, unpacked grains of calcite, siderite and limonite we were built up. In dynamic tests a greater tendency to form linear ar­ rangements characterized coat in n obtained In the ;oresence o f colloids as compared to t o o n obtained under s imilar 79 conditions without colloids. In static tests, when col­ loids were present the coatings consisted of calcite dis­ tributed between blotches of limonite and siderite. In the absence of colloids, under identical static test conditions, coatings were composed of loose grained mixtures of calcite, limonite and siderite. Generally, the size of grains for­ ming the coatings was slightly greater where colloids were present. Table 5 shows the effect of colloidal CaCC>3 in static tests. The weight gain in this table is the weight of CaCO^ coating deposited on stainless steel specimens. The weight loss is the metal lost after cleaning the corrosion products from cast iron specimens with an electric eraser. Less weight loss for cast iron specimens and more coating of stainless steel specimens resulted in tests where the water contained CaCC>3 in colloidal form than when colloids were absent, all other factors being the same* It would seem that the development of good anti­ corrosion protection requires an initial deposit of dense material which is well bonded to the metal. be but a few molecules thick. This layer need Upon such a foundation a hard and tenacious coating can then be developed. It has appeared that such initial layer should be of an adhering mixture of corrosion products and calcium carbonate. Ueither rust alone nor calcium carbonate alone possesses the desirable 60 a O ♦H CO P-T O 0 St * p St eg g 3 for Some Static Tests • 0 0 St St o o 0 —I 0 crf crf rH rH 0 0 i— l i—1 0 0 0 0 p p C O 0 P 0 P 0 p p 0 0 P P 0 0 St St o o 0 0 0 0 0 © 0 0 0 rH 0 !h 0 0 0 rH 0 0 0 •r-l •r'l pH rH St St -P -P P P C O C O ♦H ‘r-l St ., 1-CH crf crf crf crf o o erf erf P P 00 00 o o vO 1A oco crf oi crf crf 0 0 •H "H 0 m 0 o o St St 0 Ph p p 00 00 0 0 0 ft A 0 I crf 00 0 0 0 ftft o o St St 0 ft A O O A A! P P crf crf o o I —1(—) 0 0 0 0 CQ 0 © 0 rH i —[ i —1i —1 St St »SHt*r-l St •r-l *H crf crf Crf Crf P P p p 00 00 00 00 _— . i—1 p 0 0) p 0 rH A C rf Erl o *r) e o p erf * O ft p CC St ' ft (AGO 1 — 1 l—1 ;Jc oco i— 1 rH + rH O OJ OJ + H O OJ OJ p St p ft 0 0 0 0 0 0 Sh ft ft ft A O AO O crf o o crf O 0 ^.s Crfi — 1 f t rH •H Crf O P rH 0 — i1 o ft o o * H- 81 qualities of the mixture. Rust not containing CaC03 has been too porous and not sufficiently dense. Pure calcium carbonate layers, on the other hand, were soft* In the presence of CaCO^ colloids the test waters were observed to move rapidly toward equilibrium* It is well recognized that calcium carbonate remains in supersaturated solution for long periods of time unless the reaction rate of calcium and carbonate ions in forming calcium carbonate is increased by the presence of crystal nuclei or by energy in the form of turbulence, high velocity or heat. The ac­ tion of the colloidal calcium carbonate has been to accele­ rate the movement of the supersaturated solution toward equilibrium* Rapid deposition of calcium carbonate has re­ sulted in the formation of an Initial dense layer of the mixture, described above, before corrosion products had time to build up on the surface of the metal. The positive charges of colloidal particles also aided the formation of the dense, hard coatings. These parti­ cles, due to their charges, were selectively laid down on cathodic areas of the metal surface through electrodeposi­ tion. As shown in nolarization studies, It has been esta­ blished that calcium carbonate acted as a cathodic inhibi­ tor. The rate of corrosion thus was decreased because this cathodic film formation increased the polarization of cath­ odes In the corroS-Lon cells on tns surxace of one mcoal. 82 It can be concluded from the above discussion that, due to selectivity In deposition, the over-all rate of cor­ rosion was less In the presence of colloids than without such colloids, all other factors being the same. This fin­ ding was substantiated from weight loss determinations in Table 5 . With a decreased corrosion rate, calcium carbonate could interact with corrosion products in forming protective layers of the coatings. The formation of protective films on cathodic areas of the metal probably caused the forma­ tion of dlfferential-aeration cells with the shielded cath­ odes becoming the anodes of the new cells. Thus mixtures of corrosion products and calcium carbonate would be depo­ sited uniformly through the entire surface of the metal. Figure 16 shows potential-time curves for Dynamic Tests Ip and 5- Colloidal CaC03 was the only variable, all other factors being the same; Test Water 5 contained CaCO^ in colloidal form while no such colloids were present in Test Ip. Corrosion potentials were measured against the sa­ turated reference calomel electrode at the end of desired Intervals of time. The variation of potential with time during corrosion provided information of interest. Such potential variation serves to distinguish between the tendency to develop corro­ sion and the tendency to stifle corrosion. A continuously 83 Ti *+HrH vT o cJ £ m © nd o bO SP rH -p •rH £ O (1) o O 0>rtH © rH P © O h * rH o c T—H ' d t> w O cd o © •H pq O o O Q o OJ rH O O rH - Hours O V.C T ime present Colloids O CO o o CM O o I Sq.TOi\ - pQq.^tmq.’Bs o I p^pq.tioq.oj Be­ falling potential shows that the liquid is corrosive, while a steady or rising potential is noted when the liquid fa­ vors passivity. significance; The absolute level of potential is without interest attaches only to the question of whether potential rises or falls. It is to be recalled that when passivity is due to increased polarization of the cath­ ode the resulting corrosion potentials are of less noble values (higher negative values) as compared to a standard calomel electrode. With the absence of colloids in Run k, the corrosion potential of cast iron specimen changed sharply to a less noble potential in the first hour. This Initial change was followed by a rather sharp and gradual shift In the noble direction over the remainder oeriod of 120 hours* The In­ crease in negative value of the potential during the first hour can be attributed to the formation of corrosion pro­ ducts on the surface of the snecimen. The subsequent de­ crease in potential values Is probably due to removal of corrosion products from the surface of the specimen as the run was continued. The corrosion potential of Test 5 where colloids were present shows a rapid saiit en a less n o d e direction, --1“ lowed by a stead' and gradual s'dfT In the same direction for the rest of k 8 hours period. dls shift indicate'-1 that the degree of pass ivity formed on cast iron specmie^ at the 85 beginning of the test was greater than for Run !{_* Passi­ vity was maintained, and increased for the remaining period of the test. It is to be noted that CaCO^ crystals of larger than colloidal size did not have as favorable an effect on for­ mation of coating, as colloidal CaCC>3 . These crystals showed a typical calcite form under a microscope and car­ ried no electrical charge. The Effect of Momentary Excess on Coating Development In an aqueous solution, the calcium and carbonate ions present which are in excess of the solubility product con­ stant of calcium carbonate have been defined by Dye (32) as the "momentary excess." Momentary excess Is invariably less than "saturation excess," which Is the amount of calcium, carbonate precipitation teat must take place to bring a calcium-carbonate-bicarbonate-carbon dioxide system to equilibrium. ■Since the concentration of carbonate ions in a Iven solution Is a function of the pH level of that solution, high momentary excess values resulm from high pH levels. Low pH levels, conversely, produce low momentary excess values. The momentary excess level represents the driving force tending toward calcium carbonate deposition. h r 86 relationship, therefore, should exist between this excess and the rate of calcite deposition. Table 6 shows values of momentary excess and satura­ tion excess for waters which were studied in dynamic tests. In all tests shown the water contained calcium carbonate colloids. Tests were run for the period of time required to produce a definite coating, commonly one to five days. Tests 7 and 8 were made at high pH levels (high mo­ mentary excesses). The coating on Test Specimen 7 was soft limonite covered heavy, loose grained, and chalky depo­ sits of calcite. by It was apparent that there was no Inter­ mixing between limonite and calcite, especially in the layers closest to the metal. 'The coating of Test 8 consisted al­ most entirely of soft limonite, very little calcite being present. The rate of deposition of CaCOg was very high in these two tests. However, only a small part of the material was deposited on the test specimen, the remainder plating out at the bottom of the barrel reservoir and in the Tygon tubing of the apparatus. Momentary excess values for the other tests of Table 6 , were of' much lower order than for the first two high pH studies. Hard, dense coatings were developed on all speci­ mens with good to excellent oond to ire metal. "ra ins found in t oe coat ongs o _ oalcitc oeso tests •■•••]ere o j. a o ~ half the size of those found at ’'I'"h p” levels. Uniform 87 O © B rH w £ __. •h p cd > O J eo £ O •H 0 p cd p C O O O © £ P 0 0 C O © p ----- __ with XS > fcC •H Cj -p *(— } VT\ OJ rH 1—1 • VO 0 • VO 0 0 0 • --T -or VO • -d 1— 1 O O OJ OJ 0 adjusting p to w © 0 VO « xO 0• -d" rO -d • 0 • sT; xC VO. • CO PT O • O CO O • OJ OJ r^| vo vC 0 vr\ 1— 1 0 OJ pH VON OJ • O «H vC • co — • j. r• OJ k © £ O •H -P cd £ B p cd CO 00 co © 0 X! © CO O w 0 © cd O • -d vo O OJ £ £ cd P tB P P P C O * * £ cd p £ © £ 0 S 10,0, then f~) cd © 0 Pd 0 by raising cd o- developed Table 6 Relationship Between Saturation Excess, Momentary Excess, Type of Goatins Developed in Dynamic Tests © Colloids and > •r-l *0 -P Ph © CO2 . cd intermixing of calcite and corrosion products generally ap­ peared in all the tests with low momentary excess values. Figure 17 shows the corrosion potentials of the speci­ mens during some of the tests shown in Table 6 , Potential values for Tests 7 and 8 of high pH levels and high momen­ tary excess do not indicate a high level of passivity as compared to those of Tests 11 and 12 of low pH levels. The relative protective values of coatings developed on cast iron specimens in static tests in the presence of CaC03 colloids is shown in Table 7 for different levels of momentary excess. This evaluation is also based upon hard­ ness and bonding to metal. Coating on the specimen from Test 13, which was con­ ducted at a high momentary excess level of 2_|_6 • , was the poorest. This Test 13 coating was made of soft blotches of limonite with very little calcite. In Test 9 a low momen­ tary excess value of O.h.2 also produced a poor coating con­ sisting largely of soft blotches and aggregates of limonite and siderite. At a momentary excess level of 11.1 in Test 19, the coating obtained was made of calcite distributed between blotches of limonite and siderite. The limonite appeared to be relatively harder than that formed on specimens of Tests 11 and 13. Similar arrangement of calcite deposition between blotches of limonite and siderite was also obtained 89 Figure 17 - Corrosion Potentials at Different 11Momentary Excess” Values (ME = Momentary Excess) est 12 Potential to Saturated Calomel - Volts Test 11 Time ”W Hours 100 *120 90 Cm O __^ ft 0 h q _—^ ft P cd aj t> s o o 0 > •H ft Table 7 Relationship Between Momentary Excess Levels and Type of Coating Developed on Cast Iron Specimens in Static Tests ' ft i— i P 0 O ft 0 O P f-1 O 0 P. > f t CD no 0 p cd o p CO o 0 u p w 0 ft o o ft ' -CM > ft h P *rl cd P ft IP CM O o rft pH cd 0 (ft ft *rd bD .S O O -P GO P •i—i cd h CD 5h cd P ft CO £ CO 0 0 S O O X S 0 ft P D-- 1ft r-l i —I CM O o 1—I CD > O cd ft ft ft h cr CO O CO o 0 £o •rl CO cd o £ Ph K Pi Jh cd co CO CM (ft CM Cft o vr\ •iM Cd Sh ft ft 0 P< o r-t 1ft CM • O O ft• Cft CO • 0 ft • CO > 0 ft r-l CO ft •H o P CO * CD O eh hi PO i-l ft ft '!• ft i—i rH O ( ft O 91 on specimen of Test 6 * The blotches, however, were smaller in size than those formed on the specimen of Test 11, and the coating appeared slightly harder than that obtained in Test 11. The same evaluation is shown in Table & for stainless steel specimens in static tests under different momentary excess levels. In these tests, the specimens were under a uniform applied current density of 1 .031+ ma/dm^, and col­ loidal CaC03 was present in the test waters. The weight of CaC03 deposited on the cathodes and anodes of stainless steel specimens after a period of seven days is used as an index of the effect of momentary excess level. Once again, coatings on specimens obtained from tests with high levels of momentary excess were the poorest, based upon hardness and bonding to metal. Better coatings developed in the tests with lower levels of momentary ex­ cess . No relationship could be obtained between the weight of CaC03 deposited on the anodes and cathodes of stainless steel specimens and momentary excess levels. The data in­ dicate, however, that the weight deposited on the specimens was more nearly related to hardness than to momentary ex­ cess. G-enerally the higher the level of hardness the mo-e CaCC>3 deposited on both the cathode and anode of stainless steel specimens. 92 1 •H m o CD ,— . ft P h,P £ O 0 P Ti bO ft P ■P m o Table 8 Relationship between Momentary Excess Levels, Type of Coatin, Developed, and height of CaCO^ Coatings Deposited on Stainless Steel Specimens in Static Tests CD to o ft •H O T i O (D cd CD P gc O P Cd 1 •rH cvt if t 1— I o o ft) o o • o L ft ft ft o o • o o o ft —f- • e ft ft C o • o C ft lf t ~r O o O 1 ft CO ft c • ft ft ft, Hv ft o _—„ m £ O CD bC '— O 0 P CD TJ ft o 1 ft e ft ft P 'P ,P m o O to o ft •H O op P o ■^D o • o • o r ft O • O ft! CO 1 ft o • o o ft • o o o • CD P CD GJ p: O P o CD > bO -—. bO •H P bO O P CD ft P P O Ctf P O P •H P cd O O ft p p w ft o ft CD ft CD o > •H o P CD ft cd P 0 ! i—IrH >1 CD cd CD ft > ft] F !>* 1 P F cd p co P CO CD 0) g °i O x\ t —i f^ —{ o] 0 P o cd o O P 0 0 o rO '-' ft "— eft ! 1--1 CO i— 1 • Cft • ft) ft) f t ft i—) -G i­ ft o o ft O CO P ft o ft —G f ■Lft ftj • '. f t • o o i—1 • O ft- ■ft PA 1ft • 'f t i— • • OJ • o 1— I M d co o 0 Ctf P Si tP P h! (ft cc! Ph t I —r —*i t I— II P hI ft CO • 00 o ft P ft OJ ft! i—! ft! ''f t • o —1 — ft] o ft! o ft CO • CO ft • cm ft) c\' ft] 93 From the above analyses of the relationship between momentary excess levels and the coating developed in both dynamic and static tests, it was considered that the anti­ corrosion value of these coatings was, generally, in in­ verse ratio to momentary excess levels. That is, low m o ­ mentary excess levels produced superior type coatings. Low momentary excess levels with high saturation ex­ cess values, in the presence of CaCC>3 colloids, led to the formation of the best coating developed, especially in dy­ namic tests. tective. These coatings were tenacious, hard and pro­ This combination of low momentary excess and high saturation excess occurs when the test water is in the pH range of 8.2 mo 8.7 and has a momentary excess between 2.5 and 5*5* Levels of pH higher than 9.5 were definitely es­ tablished as undesirable. Effect of the Age of Colloidal Particles On the date of Deposition of CaCC>3 and Coating Development It was found that the rate of deposition of CaCO^ from supersaturated solutions containing colloidal material i-mis not only related to themomentary excess levels of the so­ lutionsbut also to the age of hue colloidal material pre­ sent . Three Identical Dynamic Tests Ik, 15, and 16 were made 9k with, the same testing water at pH of 8*3 and momentary ex­ cess value of [{-*0* Test 15 was made four days after Test lk while Test 16 was made twenty days after Test lk* The only difference between the tests lay in the age of the colloidal material in suspension* Precipitation in Test lk was 15 ppm per day, in Test 15 was k5 ppm per day, and in Test 16 was 60 ppm in six hours. Both Tests lip and 15 produced good protective coatings in one day's time. The coating of the Test lk specimen was of smoother texture and did not show the very marked ridges and valleys of the specimen's coating of Test 15* The coating of Test 15 seemed slightly harder and better bonded to the metal surface. Figure 18 shows corrosion potential-time curves for Tests lk and 15* The figure indicates that the passivity acquired by the specimen of Test 15 was still increasing at the end of one day's time, while that of Test lk had reached an almost steady level by the end of a half-hour period. Coatings from Test 16, where the rate of precipitation was high (60 ppm in 6 hours), were checked at the end of six hours and found to be very poor. The coating could be removed easily by a finger nail as the limonite was very soft. Analyses indicated that percentage calcite in the coating was much less than that of Tests lu and lp, although the rate of deposition of CaCOy was much higner. o o o O O I I S3-T°A - I©tuoiB0 poq.'eauaBs o- XBXiuoq.oj o-> •H s E-t of late of Deposition of CaCO on Coatings Developed I 13 - Effect m 0 O figure 95 96 The effect of the age of colloids on the rate of depo­ sition of CaCC>3 and consequently on the coatings developed can also be shown in Dynamic Tests 12 and 13. These two tests were also conducted at pH of b.3 and momentary excess value of ij..O. Test 13 was made 7 days after the beginning of Test 12, using the same testing water. The average pre­ cipitation in Test 12 was 10.2 ppm per day and that of Test 13 was 52.5 ppm per day. Coatings developed on the specimens at the end of five days with Test 12 and four days with Test 13 were both good, dense, and hard, with good bonding to the metal. However, the coating of Test 12, was of smoother texture, somewhat harder and better bonded to the metal than that of Test 13. Potential-time curves for Tests 12 and 13 are shown in Figure 19* Specimen potential of Test 13* where the rate of deposition of CaCC>3 was higher, shifted more rapidly to large negative values at the beginning of the tests than for the specimen of Test 12. As the tests progressed, the specimen of Test 12 acquired more protection than that of Test 13* It was evident from these and other tests that the rate of deposition of calcium carbonate had very signifi­ cant influence on the type of coatings developed. food coatings were produced only under a certain range of depo­ sition rates. 3elow or above r.hat range coatings were - Volts 97 Calomel ppt. 10,2 ppm/day Test 12 -O Q Potential To Saturated pp t ♦ 52.5^ PPm/daj 20 100 120 Figure 19 - Effect of Rate of Deposit ion of CaC03 on Coating Developed 98 poor. This optimum range of deposition rate was achieved when the test water was within the range of pH and momen­ tary excess levels recommended in the discussion 011 the in­ fluence of momentary excess levels and when the colloids were newly formed. As colloids became older and probably larger in size, the rate of deposition of CaCOy increased although the test water was maintained at the same levels of pH and momentary excess* In Test 19, the rate o r deposition increased beyond the limited range recommended and resulted In a poor coating. Effect of Dissolved Oxygen Levels on Coatings Developed Levels of dissolved oxygen other than for saturation were studied in Dynamic Tests 17, 18, 19, 20, 21, 22 and 2h. Table 9 shows dissolved oxygen levels for these studies. Analysis of the coatings developed is also shown in Table 9. Test waters for Dynamic Tests 20, 21, 22 and 2l\. were identical; the pH level was 3.3, hardness was l8o ppm and colloidal CaCO-} was present. only variable in criese Dissolved oxygen level was the four tests. -'us coatin'" developed a" low dissolved oxygen level of 0 o88 to 2.1 ppm in Test 20 was soft and flawy, chisfly ('"9 ) limcnite. The percent calcite was relatively low. -erg 99 o cA U\ H £ O CD co 50 CO tA £ •H « p P O 1 TJ u cd JA Effect of Dissolved Table Oxygen 9 on Coatings Developed tc o o n no 102 The mechanism, of formation of this coating, apparently, was different from that in effect at dissolved oxygen levels below saturation. The rate of precipitation of ferric h y ­ droxide next to the metal surface probably was too fast to permit intermixing with calcite near the metal. Figure 20 shows potential-time curves for Tests 20, 21, 22 and 2L;_* The curve of Test 20, which was studied at the lowest oxygen level of 0.86 - 2.1 ppm, indicates formation of a temporary passivity at the beginning of the test. This passivity was probably due to the precipitation of a loose mixture of ferrous hydroxide, ferric hydroxide and magnetite which was removed as rhe test progressed. As dissolved oxygen leve3_s Increased In Tests 21 and 22, potential-time curves Indicate that the passivity formed at the beginning of the tests, due to formation of ferric hydroxide, acquired more permanent form. Specimen potential of Test 2i|_, which was conducted at an oxygen level of 10 11 ppm, shifted very sharply to high negative values at the beginning of the test. This rapid shifting was followed by rapid falling and then by a slower steady fall. The sharp rise of potential at the beginning of Test 21g can be attri­ buted to the excess deposition of loose ferric hydroxide, which was then remover fro?: tne surface of the metal as the test progressed. Test 17 was studied at low oxygen level of 1.5 - 2.5 103 (D*0 . = Dissolved Oxygen) To Saturated Calomel - Volts Test waters: at pH 8.3 and hardness 180 ppm* CaCO^ for all tests Potential es Time Hours Figure 20 - Effect of Dissolved Oxygen Levels on Coatings Developed loij ppm and high momentary excess value of 3k .7 with a high rate of CaC03 deposition. The test water was at a pH of 10.25, and a hardness of 65 ppm, and contained CaC03 col­ loids. The resulting coating was uneven and irregular, con­ sisting mostly of magnetite, and covering a large part of the specimen surface. Sixty five percent of the specimen coating obtained from Test 17 was black magnetite. The insufficient supply of oxygen and high rate of CaC03 deposition prevented fur­ ther oxidation of ferrous hydroxide which, precipitated with ferric hydroxide on the metal surface. Interaction of fer­ ric and ferric hydroxides thus yielded magnetite. The coating contained a relatively small percentage of calcite after 5 days. CaC03 apparently was removed as the test progressed due to poor bonding to either iron oxides or to the metal surface. The potential-time curve of Test 17 is shown in Figure 21. The curve Indicates that the cast iron specimen began to acquire some passivity only after about kO hours had elapsed. This passivity may be attributed to precipitation of sufficient ferric hydroxide to allow some intermixing with calcite at this stage of the test. Formation of large amounts of magnetite which appeared to be somewhat hard may also have contributed to this passivity. Test IB was conducted at a pF of 6.3, a hardness of 10 5 FIgurej21 - Effect of Dissolved Oxygen Levels on Coatings Developed -1.U (D.0. — Dissolved Oxygen) est water 17: 1.1 - Volts - 1.0 -0.9 pH 8.3# hardness 200 ppm Test water 19: pH 8.3# hardness 200 ppm 0.8 Potential To Standard - - Test water 18: 1.2 Calomel - pH 10*2$, hardness ^5 ppm 1.7 - 2.7 ppm D.O 7-28 Test 18 ppm -0.5 -o.U Test 17 160 T ime - Hours 106 200 ppm, a dissolved oxygen level of 1.7 - 2.7 ppm, and in the presence of CaCO^ colloids. After four days the coating was soft and similar to that obtained from Test 21. A heavy calcite layer overlaid the coating, while limonite and sid­ erite were found next to the metal. Very little intermixing occurred between calcite and the products of corrosion. Potential-time curve of the specimen of this test, shown in Figure 21, indicates similar behavior to that of Test 21. A very high level of dissolved oxygen was used in Test 19* This test was at a pH of o.3> a hardness of 200 ppm, a dissolved oxygen level of 2.1 - 28 ppn, and in the presence of colloidal CaC0 3 . A soft coating was obtained. An excess of corrosion products was observed near the metal surface covered by a calcite layer. Obviously, the deposition of ferric hydroxide near the metal surface was too fast to permit intermixing with calcite. The distinguishing characteristic of the coating ob­ tained from Test 19 and as compared with that obtained from Test 2i| at an oxygen level of 10 - 11 ppm was one of unifor­ mity. The coating of Test 19 covered the entire surface of the specimen while that of Test 2h covered only pare of the metal surface. Figure 21 snows the potential-time curve of Test 19. The ;eneral snare of the c rve u s 2b. shown in - ‘mure 2T. similar to that for Test end "all of potential 10' values at the beginning of Test 19 was greater than the corresponding shift of potential values in Test 2) 4. This was probably due to greater precipitation of loose ferric hydroxide on the metal surface at the early stage of Test 19 and subsequent removal of these precipitates. It can be concluded from the above discussion that in order to develop a good, hard, and uniform coating, a coor­ dination should exist between the rate of deposition of CaC03 and the rate of formation and precipitation of ferric hydroxide next to the metal surface. sition of CaCOg studied, At the rates of depo­ it appeared that saturation of test water with dissolved oxygen was the optimum level, permitting thorough Intermixing of calcium carbonate and consequently resulting in the development of a favorable coating• Effect of Specimen ourface Conditions on Coatings bevelowed Two cast Iron specimens were of the flow system of Test 11. munted In the test coll One of the specimens was sand blasted while no surface c enact ion m 3 was applceo. to the second specimen ot>er t h u - standard procsd-" surface grinding with a hO-g-it diamond dressed wheel was used for all cast iron s an too study. 108 test was conducted at a pIT of 8 .1±, a hardness of 205 ppm, saturation with dissolved oxygen, and in the presence of CaCC>3 colloids. No difference in composition, hardness, or bonding to metal was noticed in the coatings developed on the two spe­ cimens. Potential-time curves revealed a slight difference in the passivity acquired by the two specimens, as shown in Figure 22. The sand blasted specimen curve slows slightly higher passivity than for the smooth metal, especially af­ ter about 80 hours had elapsed. Toughening the surface of the cast iron specimen probaoly helped the bonding of the coating to the metal surface to some degree. Two identical dynamic Tests 12 and 13 were ■'undertaken with identical test water; pH b.3> hardness 200, saturated with dissolved oxygen, and containing colloidal CaCO^. A sand blasted stainless steel specimen was used in Test 12 while Test 13 was conducted with a smooth surface stainless steel specimen. No observable deposit formed on the smooth specimen af­ ter a p riod of four days and no significant deposit was n o ­ ted on sand blasted specimen after five days. This absence of the formation of coatings on stainless steel specimens flow system without applied flfF substantiated the work o^ Strum (15). Strum has reported that practically no CafC^ was precipitated at, the srr.’ace of stainless steel snrfnce in Calomel - Volts 109 Sand blasted surface cast iron specimen 0.45 Potential To Saturated - - ___JD-© Test 11 Test water: Regular cast iron specimen pH 8 hardness 205 ppm -0 -.55 100 1120 Hours Figure 22 - Iffeet of Cast Iron Specimen Surface on Coatings Developed 11J in a five-day period even though the saturation index of the testing water was +1.05 ph unit. Figure 23 shows potential-time curves for Tests 12 and 13. The figure indicates that the sand blasted specimen possessed relatively better passivity than the smooth sur­ face specimen as the tests progressed. The roughness of the surface apparently helped to hold a very slight coating of CaC03 on the surface of the sand blasted specimen and thus caused the difference in passivity, A coating of CaCC)3 on the sand blasted stainless steel specimen was obtained in Static Test 28 without an electric current being applied. This static test was conducted at pH 0.8, hardness of 220 ppm, and in the presence of CaC03 colloids. An uneven, hard coating covered a large portion of the surface of the specimen. Crystals and aggregates of calcite tended to build up on one another, especially around the pits of the sand blasted surface. 'hie absence of a con­ tinuous flow of water and of the turbulence around the spe­ cimen in static tests created condition favorable to depo­ sition and bonding of CaC03 to the rough surface of the sand blasted specimen. The physical condition of metallic surface on cast iron specimens had much less Influence upon coating development than the environmental conditions discussed earlier in the section. The deposition of calcium carbonate on cist Iron 111 otential To Saturated Calomel - Volts Test waters for the two tests pH 8 .3 * Hardness 200 ppm OTest 12 20 16 0 f80~ 100 120 Time - Hours Figure 23 - Effect of Stainless Steel Specimen Surface on Coatings Developed 112 specimens and the lack of this deposition of stainless steel specimens In dynamic tests indicated the importance of corro­ sion products in the formation of the coatings on cast Iron specimens, Effect of Iron Oxide on Coating Development on Stainless Steel Specimens Ferric oxide was prepared by adding 2y of ferric chlo­ ride solution to a large volume of hot water. The resulting solution was then allowed 10 boil until a positively charged colloid of iron oxide was obtained. Stainless steel specimens were used in Static Test 29, which contained 3*2 ppm (as me) of this positively charged iron oxide colloid. .west 29 was conducted at pH ". 3, hard­ ness of 90 ppm and in the presence of CaCOy collo h h . "Tn applied electric current was used. It was observed that when colloidal ferric oxide was added to the test water at pH 9.b large particles o " iron oxide were formed, most of which precipitated quickly Into the bottom of the glass cylinder. These particles c-r-lor nc electric c lar ge • ron oxe.de The level periodically in the same way as h ? for pH and hardness of test Calcium *•” - m ao i istod too standard nr or tor. coatings dev sloped a h e ? a week* ~ 113 time contained 1 - 2 percent iron oxide. The coating was harder and tougher than any other coating of CaC03 developed on stainless steel specimens in this study with or without applying electric current. This test showed clearly the ef­ fect of intermixing of CaCO^ and iron oxide on the coatin': developed. Dynamic Test 2p was conducted at ph 6 .3 , hardness 200 ppm, and in the presence of CaC03 colloids. One ppm of iron oxide colloid (as he) was added to the test water. blasted stainless steel specimen was used. A sand Iron oxide col­ loids acted in the same manner as in static Test 29, growing into large particles, most of wlled.: precipitated quickly. The level of iron oxide was adjusted periodically, as in the regular procedure for pH and hardness of test water. Ho ap­ plied electric current was used. An uneven soft coating of CaC03 with 1 - 2 oxide was obtained after three day!s period. the surface of the specimen was coated. percent Iron Only mart of Iron o ide was de­ posited close to CaC03 deposits. It Is to be recalled that it 'was not possible to obtain any observable coating of CaC03 on stainless steel spec*mens (with sand olasted and regular surfaces) in dynamic Tests 12 and 13 which were conducts! under identical conditions o p V~H 8 .3 , hardness 200 ppm, and in -he presence of Cafig hut ^ no Iron oxide present. * +- n 11' Iron oxide, therefore, aided in the formation of the coating obtained In Tesc 25. hoe rate of deposition of CaGO^ was relatively large due to the presence iron oxide particles which aided the growth and precipitation of CaC03 colloids. Probably high rate of deposition of CaCC»3 with insufficient amount of iron oxide to Interact with caused the formation of soft coating. Static Tests VS. Dynamic Tests G-enerally, coatings obtained from static tests were soft and poorly Doncied to the metal. cast Iron specimens Coatings developed on in the absence of CaCOy colloids were a mixture of calcium and ferrous carbonate and Iron oxide, bonded to a porous layer of rust. In the presence of col­ loidal material, the coatings consisted of calcite distri­ buted between blotches of limonite and siderlte. Better bonded, harder, and more uniform coatings gene­ rally were obtained in dynamic tests, A linear arrangement of ridges and valleys x^ras the most prominent characteristic distinguishing dynamic tests from static tests. in most dynamic tests, calcite was found largely in 1me valleys while ridges were commonly limonite at the surface w:_to siderlte depos ited below near ins motal• -trier favorable conditions, the rid ;:es and valleys we^e o n It upon a uni- or' 115 layer of mixed calcite and corrosion products well bonded to the metal. ferric hydroxide requires oxyyen for its precipitation. In dynamic tests under saturation condition, oxyyen was In excess and uniformly available over the entire surface of the specimen. Ferric hydroxide, could then be precipitated very close to the metal in a uniform thin layer. Subsequent deposition of calcium carbonate uniformly over the entire surface would lead to thorough intermixing, resulting in a uniform protective coating. In static tests, the source of oxygen was at a dis­ tance from the metal and the main site of precipitation of ferric hydroxide was not close to the metal surface. The precipitation was then concentrated oncertain areas of the metal and blotchy coatings resulted frommany static tests. In addition, the specimens did not shew uniform deoos ition of CaCOg as In menamic tests. All dynamic tests studlem were at a linear velocity of two feet oer second^ Since dynamic and static tests conduc­ ted under Identical conditions r'auced two distinctive types of coatings It is o b v h ’u J_hat velocity nas s I'pi, icant Influence on coating demtounent. 116 COBCLUoIONo The following conclusions are drawn from the investi­ gations undertaken in this thesis: !• Calcium carbonate acted primarily as a cathodic in­ hibitor • 2. The action of calcium carbonate in developing pood protective coatings lay in its forming a physical mixture with corrosion products. 3* Coating materials developed on cast iron specimens were largely hydrous ferric oxide in the form of limonite. From p to !;.0 percent calcite was commonly present. Ciderite and magnetite were usually observed, covered by limonite and calcite. for sed the hith stainless steel specimens calcite alone nt ire coat in•". h. Better nrotection and ‘'ether bonded, harden, and ton :her coatings resulted frc i so 'shions containin'' col­ loidal CaCOg than from ident'c'l solutions of the same p.'" and hardness with no colloids present. 5. Colloidal CaCOp had a nos *-ive charge in th~- r” ran me of 6 to 11 and 6. say se re m did not %av CaCO^ crystals in suype able effect on formation of cost 117 solutions. 7. A high "momentary excess '5 level of calcium carbonate led to the formation of chalky soft coatings, Lou "momen­ tary excess" and high "saturation excess" levels of CaC03 led to the formation of tenacious and hard, protective coatings, 8 , Rate of deposition of GaCO^ from supersaturated so­ lutions containing colloidal material was Influenced not only by the "momentary excess" value but also by the age of colloids present, 9 * A dissolved oxygen level of saturation was optimum for development of good protective coatings under the con­ ditions carried out in this study. 10. High velocity flow rates were desirable in the for­ mation of hard and dura ole coatings. btatic tests produced soft coatings. 11. Surface condition of cast iron specimens had lit­ tle effect on the type of coatings developed. 12. It has been established from the investigations un­ dertaken In this study that wren the test water was satura­ ted with dissolved oxygen, of pH range from b.2 to t "momentary excess" levels of 2. a to , with ..p , flow rate of a oovt two feet per second, and containing colloidal C a 30 3 , a u r m form, dense, hard coating which was was neve lop eu o_u eu~- • . root vemeraturs. mil _ — - — ;ended to the — - -eta! - -- - 11 KEC OiClll:DAT IOITS it has oeen established that the presence of* colloidal CaC03 in supersaturated solutions and the rate of deposition of CaCC>3 have an important effect on the type of coat inns developed. Unfortunately, the rate of deposition of CaC03 was found to oe not only a function of the "momentary ex­ cess’1 level of supersaturated solution but also of the aye of colloidal particles present. It was apparent that the older and probably larger calcium carbonate colloids in­ creased the rate of depositin''. loids would be very desirable. stabilizing of these col­ The rate of deposition of calcium carbonate could then be controlled by the level of "momentary excess” alone. A flow of two feet per second velocity was found de­ sirable in producing hard protective coatin t as compared to static conditions which produced soft coatings. level of flow rate, two feet all dynamic tests. flow rates Only one >er second, was studied thro” Investigations of the effect of o f u r eight prove beneficial in producing hard durable coatin s, -v-tgiyg-'at t ''o this studg:r warn 1lei the couth he ^ tests c o n d u c t s ’' tally ' •:1 ~ 11? normally found in a water distrl;nticn system, on coatings developed should be investigated under the general condi­ tions recommended in this study for producing good protective coatings. Jud •ament of coatings upon the basis of bond to the cast iron, hardness, and toughness has been reasonably satisfac­ tory for those staidies but a better method for evaluation is des irable„ 120 BIBL10GP A PTtY (1 ) Johnson., J*, i.ar-win, T.P., and William, E.D., "The Several Forms of Calcium Carbonate," Am. J. of Sci., Uth Series 9.1: [1.73 (1916) . (2 ) Tillmans, J., "Die Chemische Untersuchun.q; Von Wasser," Verlog von“¥ilhelm Knapp, Soale ,~"~Ger. (T936) . (3) Langelier, W.F., "The Analytical Control of AntiCorrosion Water Treatment," J. Am. hater Works Assoc., 2 3 : 1500 (1933). ~ (9) Iioore, E.W., "Calculation of Chemical Dosages Acquired for the Prevention of Corrosion," J. Hew Engl. Water Works Assoc L, Sept.: 311 (1938")". ‘ ’ (5 ) Enslow, L.H., "The Continuous Stability Indicator," Water Works & Sewage, 8 6 : 169 (199-8) • (6 ) Langelier, W.F., "Chemical Equilibria in Water Treatment," J. Am. Water Works lessee., 3 8 : T b (199-3). (7) McKinney, D.S., "The Calculation of Equilibria in Dilute Water Solutions," Proc. Am. 3oc. Test in,:;, materials, 3 9 * 1191 (1939TT (8) Larson, T.E., and Buswell, A .W., "Calculation of Calcium Carbonate saturation Index and Alkalinity Interpretations," J. Am. Water .John Assoc., : 1667 (1992). (9) Ryznar, J.W., "A Hew’ Index for Determining Amount of Calcium Carbonate Scale Permed by a Water," dW Arc. Water Works Asso c., 36: 972 (193 9) . (1 0 ) D y e , J.P., "The Calculation o. a 1 kali'"- it y a n d ’ re~ Carbon Dioxide in .iatcr by the boo of Momo"ra^hm ," J* Am. ‘ Water W orks Assoc., 3s ‘ 89a (19' ' ). 121 (1 1 ) Larson, T.K., and Ainy, -i .l n, "Corrosion by Aater a Low-blow i/eloc it v ," -T. As. ./ater .lories Assoc.. S6 : 1 (195)+). (12) Raistrick, 3., "Condensed Phosphate and Corrosion Control," Chem. 3c Ind, Ho. 1 9 : Il08 (1952). (13) Evan, U.R., "The Practical Problems of Corrosion," J. Soc. Chem. Ind., 5 7 : 55 (1921). (1I-I-) Evan, U.R., "Metallic Corrosion, Passivity, and Protect i o n * E dwa r d Arnold Pub. Co., London T1937F. (13) Haase, L.W., Deposits and Protective Coat in os," Werkstoffe u. K orroslon, 5 : 198 (1952). (16) Strum, W ., "Calcium Carbonate Deposits," J. Am. Eater Works Assoc., i|_8 : 310 (1958). (17) Baylis, J.A., "Factors Other than Dissolved Oxyyen Influencing the Corrosion of Iron Pipes," Ind. Ear, h Chem., l S : 370 (1925). (1 8 ) Schikorr, D., "The Breakdown of Metals," Barth, Leipzit (19535 . (19) Haupt, H . , "Protection of Pipes by Removal of Acids From Portable Eaters," Geswndh. In ;. , 3 E : 333 (195-1)-’ (20) Camp, T.R., "Corrosiveness of Water to Metals," J. Mew Enyl. Water Works Assoc., June: 188 (195m) . (21) Schikorr, G., "iron Hydroxides Formed in Bustiny," Angry. Chem., 191* 322 (1930). (22) Girard, A., "A Detailed Study of Formation of the Different Forms of Bust," Thesis, Lille, (1935). (23) Whitney, W.3., "The Corrosion of Iron," F. Am. Chem. noc * 25 : 395 (2'0 Whliy , H .H *, eel. , "Cornos don "andbook," Tone Sons, New York (19' '') . (25) Sue H e r , F.W., "Corrosion, Causes and -revent ion," No.hniw- ii:“'r'5.-k 3553 , W53 ^55r> '"flrd ed., 1951). (26) Eliassen , A. , and Lamb, W.C. , 1,d6c\n.nisni of the Internal Corrosion of Eater Pine," J. Am. Eater ■dorks a s s o c . , ii.5: 1281 (1953)." ~~ (27) Hadley, R.F., "Microbiologleal Anaerobic Corrosion," Am. G-as Assoc. Proc., 2 2 : 761; (19LlO). (28) Wilson, R.E., "The Mechanism of the Corrosion of Iron and Steel in Natural Waters and the Calculation of Specific Rates of Corrosion," Ind. En~. A Chem., 1 > : 127 (1923). ‘ (29) Powell, S.I., Bacon, H.I., and Lill, J.P., "Corrosion Prevention by Controlled Calcium Carbonate Scale," Ind. Eng. & Chem., 3 7 : '1.2 (19'l5). (30) Dye , J .? ., "Calculation oh‘ ~'ffeet of Temperature on PT’, Free Carbon Dlorido and the Three Forms of Alkalinity," I. Am. oter Works Assoc., 1K : V — (1952). ~ (31) Hoover, C.?., "Stabilization of Lime-Softened Water," -T. Am. Water 'Works Assoc., 3 h : 1125 (192f-2) • (32) Dye, J.F., Private Communication. (33) HachVer Catalog, Hach Chemical Co., Ames, Iowa, Catalog Ho. 3 . (31)-) "Standard Methods for the Examination of 'later, Sewage, and Industrial Wastes'," American Public Health Association, IncTT ClOth ed*, 1955) (35) Von Weimarn, P.P., "The Colloidal, Amorphous, and Crystalline States of Matter," 2. Chem. Ind. Kolloid, 2 : 76 (1907). (36) Larson, T.E., and fuswell, A.M., "Wafer Softening," Ind. Eng. I Chem., 32: 132 (1950).